comp_name,description,data_type,comp_type,subtitle,EvaluationAlgorithmAbbreviation,data_sources,metric_type
Cornell Birdcall Identification,"Do you hear the birds chirping outside your window? Over 10,000 bird species occur in the world, and they can be found in nearly every environment, from untouched rainforests to suburbs and even cities. Birds play an essential role in nature. They are high up in the food chain and integrate changes occurring at lower levels. As such, birds are excellent indicators of deteriorating habitat quality and environmental pollution. However, it is often easier to hear birds than see them. With proper sound detection and classification, researchers could automatically intuit factors about an areas quality of life based on a changing bird population.
There are already many projects underway to extensively monitor birds by continuously recording natural soundscapes over long periods. However, as many living and nonliving things make noise, the analysis of these datasets is often done manually by domain experts. These analyses are painstakingly slow, and results are often incomplete. Data science may be able to assist, so researchers have turned to large crowdsourced databases of focal recordings of birds to train AI models. Unfortunately, there is a domain mismatch between the training data (short recording of individual birds) and the soundscape recordings (long recordings with often multiple species calling at the same time) used in monitoring applications. This is one of the reasons why the performance of the currently used AI models has been subpar.
To unlock the full potential of these extensive and information-rich sound archives, researchers need good machine listeners to reliably extract as much information as possible to aid data-driven conservation.
The Cornell Lab of Ornithologys Center for Conservation Bioacoustics (CCB)s mission is to collect and interpret sounds in nature. The CCB develops innovative conservation technologies to inspire and inform the conservation of wildlife and habitats globally. By partnering with the data science community, the CCB hopes to further its mission and improve the accuracy of soundscape analyses.
In this competition, you will identify a wide variety of bird vocalizations in soundscape recordings. Due to the complexity of the recordings, they contain weak labels. There might be anthropogenic sounds (e.g., airplane overflights) or other bird and non-bird (e.g., chipmunk) calls in the background, with a particular labeled bird species in the foreground. Bring your new ideas to build effective detectors and classifiers for analyzing complex soundscape recordings!
If successful, your work will help researchers better understand changes in habitat quality, levels of pollution, and the effectiveness of restoration efforts. Reliable machine listeners would also allow conservationists to deploy more recording units worldwide and would enable data-driven conservation at a scale not yet possible. The eventual conservation outcomes could greatly improve the quality of life for many living organismsbirds and human beings included.",audio,research,Build tools for bird population monitoring,MeanFScoreBeta,cornell-birdcall-identification,f_score_multiclass
TensorFlow Speech Recognition Challenge,"We might be on the verge of too many screens. It seems like everyday, new versions of common objects are re-invented with built-in wifi and bright touchscreens. A promising antidote to our screen addiction are voice interfaces.
But, for independent makers and entrepreneurs, its hard to build a simple speech detector using free, open data and code. Many voice recognition datasets require preprocessing before a neural network model can be built on them. To help with this, TensorFlow recently released the Speech Commands Datasets. It includes 65,000 one-second long utterances of 30 short words, by thousands of different people.
In this competition, youre challenged to use the Speech Commands Dataset to build an algorithm that understands simple spoken commands. By improving the recognition accuracy of open-sourced voice interface tools, we can improve product effectiveness and their accessibility.",audio,featured,Can you build an algorithm that understands simple speech commands?,CategorizationAccuracy,tensorflow-speech-recognition-challenge,categorization_accuracy
Multi-label Bird Species Classification - NIPS 2013,"The Neural Information Processing Scaled for Bioacoustics (NIPS4B) bird song competition asks participants to identify which of 87 sound classes of birds and their ecosystem are present in 1000 continuous wild recordings from different places in Provence, France. The data is provided by the BIOTOPE society, which maintains the largest collection of wild recordings of birds in Europe. This challenge is a more complex task than the previous ICML4B challenge, in which 77 teams participated (see proceedings at sabiod.org).
For more information about the Neural Information Processing Scaled for Bioacoustics workshop, please visit the official site.
Organizers
Pr. H. Glotin - Institut Universitaire de France, CNRS LSIS and USTV, glotin@univ-tln.fr
O. Dufour - CNRS LSIS, FR Dr. Y. Bas - BIOTOPE, FR",audio,research,Identify which of 87 classes of birds and amphibians are present into 1000 continuous wild sound recordings,AUC,multi-label-bird-species-classification-nips-2013,auc
Rainforest Connection Species Audio Detection,"Who doesnt enjoy the morning chirp of a bird or a frogs evening croak? Animals bring more than sweet songs and natural ambience to the world. The presence of rainforest species is a good indicator of the impact of climate change and habitat loss. As its easier to hear these species than see them, its important to use acoustic technologies that can work on a global scale. Real-time information, such as provided through machine learning techniques, could enable early-stage detection of human impacts on the environment. This result could drive more effective conservation management decisions.
Traditional methods of assessing the diversity and abundance of species are costly and limited in space and time. And while automatic acoustic identification via deep learning has been successful, models require a large number of training samples per species. This limits applicability to rarer species, which are central to conservation efforts. Thus, methods to automate high-accuracy species detection in noisy soundscapes with limited training data are the solution.
Rainforest Connection (RFCx) created the worlds first scalable, real-time monitoring system for protecting and studying remote ecosystems. Unlike visual-based tracking systems like drones or satellites, RFCx relies on acoustic sensors that monitor the ecosystem soundscape at selected locations year round. RFCx technology has advanced to support a comprehensive biodiversity monitoring program that allows local partners to measure progress of wildlife restoration and recovery through principles of adaptive management. The RFCx monitoring platform also has the capacity to create convolutional neural network (CNN) models for analysis.
In this competition, youll automate the detection of bird and frog species in tropical soundscape recordings. Youll create your models with limited, acoustically complex training data. Rich in more than bird and frog noises, expect to hear an insect or two, which your model will need to filter out.
If successful, youll have a hand in a rapidly expanding field of science: the development of automated eco-acoustic monitoring systems. The resulting real-time information could enable earlier detection of human environmental impacts, making environmental conservation more swift and effective.",audio,research,Automate the detection of bird and frog species in a tropical soundscape,WeightedLabelRankingAveragePrecision,rainforest-connection-species-audio-detection,reconstruction
Pitch estimation and voicing detection,"Description
Select at least two algorithms for pitch estimation (preprocessing, frequency estimation and/or postprocessing) and voicing detection.
Implement the selected methods as a kaggle python kernel or in any language and compare their performance on the test database. You can use the provided baseline kernel as starting point or reference. The baseline kernel uses a simple algorithm based on the autocorrelation to compute the pitch, without any preprocessing or post-processing.
Try to improve the results with the use of standard pre- and post-processing methods, new algorithms, a combination of systems, parameter tuning or machine learning algorithms.
Optionally, you can use the FDA-UE and PTDB-TUB databases for tuning and training.
Report the results of the assignment using a 4-pages paper format. You can use, for instance, the templates in http://www.icassp2016.com/papers/PaperKit.html#Templates. In the report you have to briefly describe the selected algorithms and initial source code including the corresponding references. Then you have to mention your experiments or original contributions and the obtained results (Detailed results on the FDA-UE and public score on the Leaderboard).
Upload the complete source code as a kaggle kernel o to a git repository (as github) and provide a link to it in the report.",audio,inclass,Analysis of basic properties of the speech signal: voicing and pitch,rmse,pitch-estimation-and-voicing-detection,mse
CIFAR-10 - Object Recognition in Images,"CIFAR-10  is an established computer-vision dataset used for object recognition. It is a subset of the 80 million tiny images dataset and consists of 60,000 32x32 color images containing one of 10 object classes, with 6000 images per class. It was collected by Alex Krizhevsky, Vinod Nair, and Geoffrey Hinton.
Kaggle is hosting a CIFAR-10 leaderboard for the machine learning community to use for fun and practice. You can see how your approach compares to the latest research methods on Rodrigo Benensons classification results page.
Please cite this technical report if you use this dataset: Learning Multiple Layers of Features from Tiny Images, Alex Krizhevsky, 2009.",image data,playground,"Identify the subject of 60,000 labeled images",CategorizationAccuracy,cifar-10,categorization_accuracy
AI for Clinical Data Analytics HW2,"Homework 2
Please check the slide for detail.",image data,inclass,Computed Tomography Lung Tumor Segmentation Task,dice,ai-for-clinical-data-analytics-hw2,segmentation
Characters classification,"Overview
The dataset consists of 332987 Chinese characters (1K classes). Train the best classifier based on cnn to outperform other participants.",image data,inclass,Chinese сharacters classification problem,categorizationaccuracy,characters-classification,categorization_accuracy
Aesthetic Visual Analysis,The evaluation metric for this competition is Accuracy.,image data,inclass,Image classification by quality,categorizationaccuracy,aesthetic-visual-analysis,categorization_accuracy
ALASKA2 Image Steganalysis,"That file you downloaded may contain hidden messages that arent part of its regular contents. The same technology employed for digital watermarking is also misused by crime rings. Law enforcement must now use steganalysis to detect these messages as part of their investigations. Machine learning is an important tool in the discovery of this secret data.
Current methods produce unreliable results, raising false alarms. One reason for inaccuracy is the many different devices and processing combinations. Yet, detection models are trained on a homogeneous dataset. To increase accuracy, researchers must put data hidden within digital images into the wild (hence the name ALASKA) to mimic real world conditions.
In the competition, youll create an efficient and reliable method to detect secret data hidden within innocuous-seeming digital images. Rather than limiting the data source, these images have been acquired with as many as 50 different cameras (from smartphone to full-format high end) and processed in different fashions. Successful entries will include robust detection algorithms with minimal false positives.
The IEEE WIFS (Workshop on Information Forensics and Security) is eager to make this happen again, as a follow up to the ALASKA#1 Challenge. WIFS is an annual event where researchers gather to discuss emerging challenges, exchange fresh ideas, and share state-of-the-art results and technical expertise in the areas of information security and forensics. WIFS has teamed up with Troyes University of Technology, CRIStAL Lab, Lille University, and CNRS to enable more accurate steganalysis.
Law enforcement officers need better methods to combat criminals using hidden messages. The data science community and other researchers can help with better automated detection. More accurate methods could help catch criminals whose communications are hidden in plain sight.
The challenge is organized by Rmi COGRANNE (UTT), Patrick BAS (CRIStAL / CNRS) and Quentin Giboulot (UTT) ; in addition to Kaggle, we have been greatly helped by the following sponsors:",image data,research,Detect secret data hidden within digital images,WeightedAUC,alaska2-image-steganalysis,auc
Northeastern SMILE Lab - Recognizing Faces in the Wild,"Do you have your fathers nose?
Blood relatives often share facial features. Now researchers at Northeastern University want to improve their algorithm for facial image classification to bridge the gap between research and other familial markers like DNA results. That will be your challenge in this new Kaggle competition.
An automatic kinship classifier has been in the works at Northeastern since 2010. Yet this technology remains largely unseen in practice for a couple of reasons:
1. Existing image databases for kinship recognition tasks arent large enough to capture and reflect the true data distributions of the families of the world.
2. Many hidden factors affect familial facial relationships, so a more discriminant model is needed than the computer vision algorithms used most often for higher-level categorizations (e.g. facial recognition or object classification).
In this competition, youll help researchers build a more complex model by determining if two people are blood-related based solely on images of their faces. If you think you can get it ""on the nose,"" this competition is for you.
The SMILE Lab at Northeastern focuses on the frontier research of applied machine learning, social media analytics, human-computer interaction, and high-level image and video understanding. Their research is driven by the explosion of diverse multimedia from the Internet, including both personal and publicly-available photos and videos. They start by treating fundamental theory from learning algorithms as the soul of machine intelligence and arm it with visual perception.",image data,playground,Can you determine if two individuals are related?,AUC,northeastern-smile-lab-recognizing-faces-in-the-wild,auc
Recursion Cellular Image Classification,"The cost of some drugs and medical treatments has risen so high in recent years that many patients are having to go without. You can help with a classification project that could make researchers more efficient.
One of the more surprising reasons behind the cost is how long it takes to bring new treatments to market. Despite improvements in technology and science, research and development continues to lag. In fact, finding new treatments takes, on average, more than 10 years and costs hundreds of millions of dollars.
Recursion Pharmaceuticals, creators of the industrys largest dataset of biological images, generated entirely in-house, believes AI has the potential to dramatically improve and expedite the drug discovery process. More specifically, your efforts could help them understand how drugs interact with human cells.
This competition will have you disentangling experimental noise from real biological signals. Your entry will classify images of cells under one of 1,108 different genetic perturbations. You can help eliminate the noise introduced by technical execution and environmental variation between experiments.
If successful, you could dramatically improve the industrys ability to model cellular images according to their relevant biology. In turn, applying AI could greatly decrease the cost of treatments, and ensure these treatments get to patients faster.
This competition is a part of the NeurIPS 2019 competition track. Winners will be invited to contribute their solutions towards the workshop presentation.
Acknowledgments
Thank you to the following sponsors & supporters of this competition:
Google Cloud: Google Cloud is widely recognized as a global leader in delivering a secure, open and intelligent enterprise cloud platform. Our technology is built on Googles private network and is the product of nearly 20 years of innovation in security, network architecture, collaboration, artificial intelligence and open source software. We offer a simply engineered set of tools and unparalleled technology across Google Cloud Platform and G Suite that help bring people, insights and ideas together. Customers across more than 150 countries trust Google Cloud to modernize their computing environment for todays digital world.
DoiT: You have the cloud and we have your back. For nearly a decade, weve been helping businesses build and scale cloud solutions with our world-class cloud engineering support. We help our customers with technical support and consulting on building and operating complex large-scale distributed systems, developing better machine learning models and setting up big data solutions using Google Cloud, Amazon AWS and Microsoft Azure.
NVIDIA: NVIDIAs (NASDAQ: NVDA) invention of the GPU in 1999 sparked the growth of the PC gaming market, redefined modern computer graphics and revolutionized parallel computing. More recently, GPU deep learning ignited modern AI  the next era of computing  with the GPU acting as the brain of computers, robots and self-driving cars that can perceive and understand the world. More information at http://nvidianews.nvidia.com.
Lambda: Lambda provides Deep Learning workstations, servers, and GPU cloud services. Lambda Deep Learning infrastructure is used by the worlds leading AI research & development organizations including Apple, Microsoft, MIT, Stanford, and the US Government. To learn more, visit www.lambdalabs.com.",image data,research,CellSignal: Disentangling biological signal from experimental noise in cellular images,CategorizationAccuracy,recursion-cellular-image-classification,categorization_accuracy
Kannada MNIST,"Bored of MNIST?
The goal of this competition is to provide a simple extension to the classic MNIST competition were all familiar with. Instead of using Arabic numerals, it uses a recently-released dataset of Kannada digits.
Kannada is a language spoken predominantly by people of Karnataka in southwestern India. The language has roughly 45 million native speakers and is written using the Kannada script. Wikipedia
This competition uses the same format as the MNIST competition in terms of how the data is structured, but its different in that it is a synchronous re-run Kernels competition. You write your code in a Kaggle Notebook, and when you submit the results, your code is scored on both the public test set, as well as a private (unseen) test set.
Technical Information
All details of the dataset curation has been captured in the paper titled: Prabhu, Vinay Uday. ""Kannada-MNIST: A new handwritten digits dataset for the Kannada language."" arXiv preprint arXiv:1908.01242 (2019)
The github repo of the author can be found here.
On the originally-posted dataset, the author suggests some interesting questions you may be interested in exploring. Please note, although this dataset has been released in full, the purpose of this competition is for practice, not to find the labels to submit a perfect score.
In addition to the main dataset, the author also disseminated an additional real world handwritten dataset (with 10k images), termed as the Dig-MNIST dataset that can serve as an out-of-domain test dataset. It was created with the help of volunteers that were non-native users of the language, authored on a smaller sheet and scanned with different scanner settings compared to the main dataset. This dig-MNIST dataset serves as a more difficult test-set (An accuracy of 76.1% was reported in the paper cited above) and achieving ~98+% accuracy on this test dataset would be rather commendable.
Acknowledgments
Kaggle thanks Vinay Prabhu for providing this interesting dataset for a Playground competition.
Image reference: https://www.researchgate.net/figure/speech-for-Kannada-numbers_fig2_313113588",image data,playground,MNIST like datatset for Kannada handwritten digits,CategorizationAccuracy,kannada-mnist,categorization_accuracy
PLAsTiCC Astronomical Classification,"Help some of the worlds leading astronomers grasp the deepest properties of the universe.
The human eye has been the arbiter for the classification of astronomical sources in the night sky for hundreds of years. But a new facility -- the Large Synoptic Survey Telescope (LSST) -- is about to revolutionize the field, discovering 10 to 100 times more astronomical sources that vary in the night sky than weve ever known. Some of these sources will be completely unprecedented!
The Photometric LSST Astronomical Time-Series Classification Challenge (PLAsTiCC) asks Kagglers to help prepare to classify the data from this new survey. Competitors will classify astronomical sources that vary with time into different classes, scaling from a small training set to a very large test set of the type the LSST will discover.
More background information is available here.
Acknowledgements
PLAsTiCC is funded through LSST Corporation Grant Award # 2017-03 and administered by the University of Toronto. Financial support for LSST comes from the National Science Foundation (NSF) through Cooperative Agreement No. 1258333, the Department of Energy (DOE) Office of Science under Contract No. DE-AC02-76SF00515, and private funding raised by the LSST Corporation. The NSF-funded LSST Project Office for construction was established as an operating center under management of the Association of Universities for Research in Astronomy (AURA). The DOE-funded effort to build the LSST camera is managed by the SLAC National Accelerator Laboratory (SLAC).
The National Science Foundation (NSF) is an independent federal agency created by Congress in 1950 to promote the progress of science. NSF supports basic research and people to create knowledge that transforms the future.




Photo Credit: M. Park/Inigo Films/LSST/AURA/NSF",image data,featured,Can you help make sense of the Universe?,WeightedMulticlassLoss,plasticc-astronomical-classification,multiclass_log_loss
Aerial Cactus Identification,"To assess the impact of climate change on Earths flora and fauna, it is vital to quantify how human activities such as logging, mining, and agriculture are impacting our protected natural areas. Researchers in Mexico have created the VIGIA project, which aims to build a system for autonomous surveillance of protected areas. A first step in such an effort is the ability to recognize the vegetation inside the protected areas. In this competition, you are tasked with creation of an algorithm that can identify a specific type of cactus in aerial imagery.
This is a kernels-only competition, meaning you must submit predictions using Kaggle Kernels. Read the basics here.
Acknowledgments
Kaggle is hosting this competition for the machine learning community to use for fun and practice. The original version of this data can be found here, with details in the following paper:
Efren Lpez-Jimnez, Juan Irving Vasquez-Gomez, Miguel Angel Sanchez-Acevedo, Juan Carlos Herrera-Lozada, Abril Valeria Uriarte-Arcia, Columnar Cactus Recognition in Aerial Images using a Deep Learning Approach. Ecological Informatics. 2019.
Acknowledgements to Consejo Nacional de Ciencia y Tecnologa. Project ctedra 1507. Instituto Politcnico Nacional. Universidad de la Caada. Contributors: Eduardo Armas Garca, Rafael Cano Martnez and Luis Cresencio Mota Carrera. J.I. Vasquez-Gomez, JC. Herrera Lozada. Abril Uriarte, Miguel Sanchez.",image data,playground,Determine whether an image contains a columnar cactus,AUC,aerial-cactus-identification,auc
APTOS 2019 Blindness Detection,"Imagine being able to detect blindness before it happened.
Millions of people suffer from diabetic retinopathy, the leading cause of blindness among working aged adults. Aravind Eye Hospital in India hopes to detect and prevent this disease among people living in rural areas where medical screening is difficult to conduct. Successful entries in this competition will improve the hospitals ability to identify potential patients. Further, the solutions will be spread to other Ophthalmologists through the 4th Asia Pacific Tele-Ophthalmology Society (APTOS) Symposium
Currently, Aravind technicians travel to these rural areas to capture images and then rely on highly trained doctors to review the images and provide diagnosis. Their goal is to scale their efforts through technology; to gain the ability to automatically screen images for disease and provide information on how severe the condition may be.
In this synchronous Kernels-only competition, youll build a machine learning model to speed up disease detection. Youll work with thousands of images collected in rural areas to help identify diabetic retinopathy automatically. If successful, you will not only help to prevent lifelong blindness, but these models may be used to detect other sorts of diseases in the future, like glaucoma and macular degeneration.
Get started today!",image data,featured,Detect diabetic retinopathy to stop blindness before it's too late ,QuadraticWeightedKappa,aptos-2019-blindness-detection,correlation
Bengali.AI Handwritten Grapheme Classification,"Challenge and dataset summary available at https://arxiv.org/abs/2010.00170
Bengali is the 5th most spoken language in the world with hundreds of million of speakers. Its the official language of Bangladesh and the second most spoken language in India. Considering its reach, theres significant business and educational interest in developing AI that can optically recognize images of the language handwritten. This challenge hopes to improve on approaches to Bengali recognition.
Optical character recognition is particularly challenging for Bengali. While Bengali has 49 letters (to be more specific 11 vowels and 38 consonants) in its alphabet, there are also 18 potential diacritics, or accents. This means that there are many more graphemes, or the smallest units in a written language. The added complexity results in ~13,000 different grapheme variations (compared to Englishs 250 graphemic units).
Bangladesh-based non-profit Bengali.AI is focused on helping to solve this problem. They build and release crowdsourced, metadata-rich datasets and open source them through research competitions. Through this work, Bengali.AI hopes to democratize and accelerate research in Bengali language technologies and to promote machine learning education.
For this competition, youre given the image of a handwritten Bengali grapheme and are challenged to separately classify three constituent elements in the image: grapheme root, vowel diacritics, and consonant diacritics.
By participating in the competition, youll hopefully accelerate Bengali handwritten optical character recognition research and help enable the digitalization of educational resources. Moreover, the methods introduced in the competition will also empower cousin languages in the Indian subcontinent.
Acknowledgements:
Apurba: Apurba is the exclusive sponsor of Bengali.AI for this competition. Apurba Technologies Inc. is founded by a group of technology veterans who have been working at the cutting edge of software development in Silicon Valley for many years. Apart from its many ventures, Apurba is a pioneer in Bengali NLP research today and is accelerating AI research in Bangladesh through its contributions.
Intelligent Machines Limited: Intelligent Machines Limited is the technical partner of Bengali.AI for this competition and is providing compute support to Bangladeshi students. IML is an Artificial Intelligence and Advanced Analytics startup offering customized solutions to businesses in Bangladesh. IML believes in the strength of Bangladeshi talented resources and in the possibility of a far greater and developed Bangladesh in the coming days.",image data,research,Classify the components of handwritten Bengali,WeightedCategorizationAccuracy,bengali.ai-handwritten-grapheme-classification,categorization_accuracy
Cassava Leaf Disease Classification,"As the second-largest provider of carbohydrates in Africa, cassava is a key food security crop grown by smallholder farmers because it can withstand harsh conditions. At least 80% of household farms in Sub-Saharan Africa grow this starchy root, but viral diseases are major sources of poor yields. With the help of data science, it may be possible to identify common diseases so they can be treated.
Existing methods of disease detection require farmers to solicit the help of government-funded agricultural experts to visually inspect and diagnose the plants. This suffers from being labor-intensive, low-supply and costly. As an added challenge, effective solutions for farmers must perform well under significant constraints, since African farmers may only have access to mobile-quality cameras with low-bandwidth.
In this competition, we introduce a dataset of 21,367 labeled images collected during a regular survey in Uganda. Most images were crowdsourced from farmers taking photos of their gardens, and annotated by experts at the National Crops Resources Research Institute (NaCRRI) in collaboration with the AI lab at Makerere University, Kampala. This is in a format that most realistically represents what farmers would need to diagnose in real life.
Your task is to classify each cassava image into four disease categories or a fifth category indicating a healthy leaf. With your help, farmers may be able to quickly identify diseased plants, potentially saving their crops before they inflict irreparable damage.
Recommended Tutorial
We highly recommend Jesse Mostipaks Getting Started Tutorial that walks you through making your very first submission step by step.
Acknowledgements
The Makerere Artificial Intelligence (AI) Lab is an AI and Data Science research group based at Makerere University in Uganda. The lab specializes in the application of artificial intelligence and data science - including for example, methods from machine learning, computer vision and predictive analytics to problems in the developing world. Their mission is: To advance Artificial Intelligence research to solve real-world challenges.""
We thank the different experts and collaborators from National Crops Resources Research Institute (NaCRRI) for assisting in preparing this dataset.

This is a Code Competition. Refer to Code Requirements for details.",image data,research,Identify the type of disease present on a Cassava Leaf image,CategorizationAccuracy,cassava-leaf-disease-classification,categorization_accuracy
Galaxy Zoo - The Galaxy Challenge,"Understanding how and why we are here is one of the fundamental questions for the human race. Part of the answer to this question lies in the origins of galaxies, such as our own Milky Way. Yet questions remain about how the Milky Way (or any of the other ~100 billion galaxies in our Universe) was formed and has evolved. Galaxies come in all shapes, sizes and colors: from beautiful spirals to huge ellipticals. Understanding the distribution, location and types of galaxies as a function of shape, size, and color are critical pieces for solving this puzzle.
The Whirlpool Galaxy (M51). Credit: NASA and European Space Agency
With each passing day telescopes around and above the Earth capture more and more images of distant galaxies. As better and bigger telescopes continue to collect these images, the datasets begin to explode in size. In order to better understand how the different shapes (or morphologies) of galaxies relate to the physics that create them, such images need to be sorted and classified. Kaggle has teamed up with Galaxy Zoo and Winton Capital to produce the Galaxy Challenge, where participants will help classify galaxies into categories.
  Image Credit: ESA/Hubble & NASA
Galaxies in this set have already been classified once through the help of hundreds of thousands of volunteers, who collectively classified the shapes of these images by eye in a successful citizen science crowdsourcing project. However, this approach becomes less feasible as data sets grow to contain of hundreds of millions (or even billions) of galaxies. Thats where you come in.
This competition asks you to analyze the JPG images of galaxies to find automated metrics that reproduce the probability distributions derived from human classifications. For each galaxy, determine the probability that it belongs in a particular class. Can you write an algorithm that behaves as well as the crowd does?
Contributors: D. Harvey, C. Lintott, T. Kitching, P. Marshall, K. Willett, Galaxy Zoo 
Acknowledgments
The Contributors and the rest of the Galaxy Zoo and Kaggle teams would like to say a big thank you to Winton Capital for helping make this happen. Without their support, we would have not been able to make this competition go ahead.",image data,research,Classify the morphologies of distant galaxies in our Universe,RMSE,galaxy-zoo-the-galaxy-challenge,mse
Facial Keypoints Detection,"The objective of this task is to predict keypoint positions on face images. This can be used as a building block in several applications, such as:
tracking faces in images and video
analysing facial expressions
detecting dysmorphic facial signs for medical diagnosis
biometrics / face recognition
Detecing facial keypoints is a very challenging problem.  Facial features vary greatly from one individual to another, and even for a single individual, there is a large amount of variation due to 3D pose, size, position, viewing angle, and illumination conditions. Computer vision research has come a long way in addressing these difficulties, but there remain many opportunities for improvement.
This getting-started competition provides a benchmark data set and an R tutorial to get you going on analysing face images. Get started with R >>
Acknowledgements
The data set for this competition was graciously provided by Dr. Yoshua Bengio of the University of Montreal. James Petterson.",image data,getting started,Detect the location of keypoints on face images,RMSE,facial-keypoints-detection,mse
Dogs vs. Cats Redux: Kernels Edition,"In 2013, we hosted one of our favorite for-fun competitions:  Dogs vs. Cats. Much has since changed in the machine learning landscape, particularly in deep learning and image analysis. Back then, a tensor flow was the diffusion of the creamer in a bored mathematicians cup of coffee. Now, even the cucumber farmers are neural netting their way to a bounty.
Much has changed at Kaggle as well. Our online coding environment Kernels didnt exist in 2013, and so it was that we approached sharing by scratching primitive glpyhs on cave walls with sticks and sharp objects. No more. Now, Kernels have taken over as the way to share code on Kaggle. IPython is out and Jupyter Notebook is in. We even have TensorFlow. What more could a data scientist ask for? But seriously, what more? Pull requests welcome.
We are excited to bring back the infamous Dogs vs. Cats classification problem as a playground competition with kernels enabled. Although modern techniques may make light of this once-difficult problem, it is through practice of new techniques on old datasets that we will make light of machine learnings future challenges.",image data,playground,Distinguish images of dogs from cats,LogLoss,dogs-vs.-cats-redux:-kernels-edition,log_loss
Gallivanters,"After their gaming startup PlayDoom couldnt garner funding in the Entrepreneurship Summit, the team of Bhadage, Bansal and Pandey decided to give up on their startup and go on a world trip to follow their passion of numismatics (collection of coins). They travel to various countries and collect coins of various denominations to add to their collection.
Now, theyve hired a data scientist Madhup to create a tool to predict countries from their coin currencies, by making use of their marvellous coin collection.
Can you help Madhup with the tool?
Note - If you are applying CNN, be ready with keras scratch code as well as transfer learning code.
Data has been updated. So kindly re-download the data.",image data,inclass,Help data scientist to create a tool to predict countries from their coin currencies,categorizationaccuracy,gallivanters,categorization_accuracy
TReNDS Neuroimaging,"Human brain research is among the most complex areas of study for scientists. We know that age and other factors can affect its function and structure, but more research is needed into what specifically occurs within the brain. With much of the research using MRI scans, data scientists are well positioned to support future insights. In particular, neuroimaging specialists look for measurable markers of behavior, health, or disorder to help identify relevant brain regions and their contribution to typical or symptomatic effects.
In this competition, you will predict multiple assessments plus age from multimodal brain MRI features. You will be working from existing results from other data scientists, doing the important work of validating the utility of multimodal features in a normative population of unaffected subjects. Due to the complexity of the brain and differences between scanners, generalized approaches will be essential to effectively propel multimodal neuroimaging research forward.
The Tri-Institutional Georgia State University/Georgia Institute of Technology/Emory University Center for Translational Research in Neuroimaging and Data Science (TReNDS) leverages advanced brain imaging to promote research into brain health. The organization is focused on developing, applying and sharing advanced analytic approaches and neuroinformatics tools. Among its software projects are the GIFT and FIT neuroimaging toolboxes, the COINS data management system, and the COINSTAC toolkit for federated learning, all aimed at supporting data scientists and other neuroimaging researchers.
Making the leap from research to clinical application is particularly difficult in brain health. In order to translate to clinical settings, research findings have to be reproduced consistently and validated in out-of-sample instances. The problem is particularly well-suited for data science, but current approaches typically do not generalize well. With this large dataset and competition, your efforts could directly address an important area of brain research.
Acknowledgments",image data,research,"Multiscanner normative age and assessments prediction with brain function, structure, and connectivity",WMAE,trends-neuroimaging,mae
Traffic signs classification,"     .      RTSD
      48x48 .  66 , 66  .      ,       .


!",image data,inclass,Russian road traffic signs classification,categorizationaccuracy,traffic-signs-classification,categorization_accuracy
Fashion MNIST-ITBA-LAB 2020,Competencia Laboratorio de Deep Learning ITBA,image data,inclass,Clasificar las imagenes entre las 10 categorías,categorizationaccuracy,fashion-mnist-itba-lab-2020,categorization_accuracy
Diabetic Retinopathy Detection,"Diabetic retinopathy is the leading cause of blindness in the working-age population of the developed world. It is estimated to affect over 93 million people.
The US Center for Disease Control and Prevention estimates that 29.1 million people in the US have diabetes and the World Health Organization estimates that 347 million people have the disease worldwide. Diabetic Retinopathy (DR) is an eye disease associated with long-standing diabetes. Around 40% to 45% of Americans with diabetes have some stage of the disease. Progression to vision impairment can be slowed or averted if DR is detected in time, however this can be difficult as the disease often shows few symptoms until it is too late to provide effective treatment.
Currently, detecting DR is a time-consuming and manual process that requires a trained clinician to examine and evaluate digital color fundus photographs of the retina. By the time human readers submit their reviews, often a day or two later, the delayed results lead to lost follow up, miscommunication, and delayed treatment.
Clinicians can identify DR by the presence of lesions associated with the vascular abnormalities caused by the disease. While this approach is effective, its resource demands are high. The expertise and equipment required are often lacking in areas where the rate of diabetes in local populations is high and DR detection is most needed. As the number of individuals with diabetes continues to grow, the infrastructure needed to prevent blindness due to DR will become even more insufficient.
The need for a comprehensive and automated method of DR screening has long been recognized, and previous efforts have made good progress using image classification, pattern recognition, and machine learning. With color fundus photography as input, the goal of this competition is to push an automated detection system to the limit of what is possible  ideally resulting in models with realistic clinical potential. The winning models will be open sourced to maximize the impact such a model can have on improving DR detection.
Acknowledgements
This competition is sponsored by the California Healthcare Foundation.
Retinal images were provided by EyePACS, a free platform for retinopathy screening.",image data,featured,Identify signs of diabetic retinopathy in eye images,QuadraticWeightedKappa,diabetic-retinopathy-detection,correlation
Digit Recognizer,"Start here if...
You have some experience with R or Python and machine learning basics, but youre new to computer vision. This competition is the perfect introduction to techniques like neural networks using a classic dataset including pre-extracted features.
Competition Description
MNIST (""Modified National Institute of Standards and Technology"") is the de facto hello world dataset of computer vision. Since its release in 1999, this classic dataset of handwritten images has served as the basis for benchmarking classification algorithms. As new machine learning techniques emerge, MNIST remains a reliable resource for researchers and learners alike.
In this competition, your goal is to correctly identify digits from a dataset of tens of thousands of handwritten images. Weve curated a set of tutorial-style kernels which cover everything from regression to neural networks. We encourage you to experiment with different algorithms to learn first-hand what works well and how techniques compare.
Practice Skills
Computer vision fundamentals including simple neural networks
Classification methods such as SVM and K-nearest neighbors
Acknowledgements 
More details about the dataset, including algorithms that have been tried on it and their levels of success, can be found at http://yann.lecun.com/exdb/mnist/index.html. The dataset is made available under a Creative Commons Attribution-Share Alike 3.0 license.",image data,getting started,Learn computer vision fundamentals with the famous MNIST data,categorizationaccuracy,digit-recognizer,categorization_accuracy
Dog Breed Identification,"Whos a good dog? Who likes ear scratches? Well, it seems those fancy deep neural networks dont have all the answers. However, maybe they can answer that ubiquitous question we all ask when meeting a four-legged stranger: what kind of good pup is that?
In this playground competition, you are provided a strictly canine subset of ImageNet in order to practice fine-grained image categorization. How well you can tell your Norfolk Terriers from your Norwich Terriers? With 120 breeds of dogs and a limited number training images per class, you might find the problem more, err, ruff than you anticipated.
Acknowledgments
We extend our gratitude to the creators of the Stanford Dogs Dataset for making this competition possible: Aditya Khosla, Nityananda Jayadevaprakash, Bangpeng Yao, and Fei-Fei Li.",image data,playground,Determine the breed of a dog in an image,MulticlassLoss,dog-breed-identification,multiclass_log_loss
Dogs vs. Cats,"In this competition, youll write an algorithm to classify whether images contain either a dog or a cat.  This is easy for humans, dogs, and cats. Your computer will find it a bit more difficult.
Deep Blue beat Kasparov at chess in 1997.
Watson beat the brightest trivia minds at Jeopardy in 2011.
Can you tell Fido from Mittens in 2013?
The Asirra data set
Web services are often protected with a challenge thats supposed to be easy for people to solve, but difficult for computers. Such a challenge is often called a CAPTCHA (Completely Automated Public Turing test to tell Computers and Humans Apart) or HIP (Human Interactive Proof). HIPs are used for many purposes, such as to reduce email and blog spam and prevent brute-force attacks on web site passwords.
Asirra (Animal Species Image Recognition for Restricting Access) is a HIP that works by asking users to identify photographs of cats and dogs. This task is difficult for computers, but studies have shown that people can accomplish it quickly and accurately. Many even think its fun! Here is an example of the Asirra interface:
Asirra is unique because of its partnership with Petfinder.com, the worlds largest site devoted to finding homes for homeless pets. Theyve provided Microsoft Research with over three million images of cats and dogs, manually classified by people at thousands of animal shelters across the United States. Kaggle is fortunate to offer a subset of this data for fun and research. 
Image recognition attacks
While random guessing is the easiest form of attack, various forms of image recognition can allow an attacker to make guesses that are better than random. There is enormous diversity in the photo database (a wide variety of backgrounds, angles, poses, lighting, etc.), making accurate automatic classification difficult. In an informal poll conducted many years ago, computer vision experts posited that a classifier with better than 60% accuracy would be difficult without a major advance in the state of the art. For reference, a 60% classifier improves the guessing probability of a 12-image HIP from 1/4096 to 1/459.
State of the art
The current literature suggests machine classifiers can score above 80% accuracy on this task [1]. Therfore, Asirra is no longer considered safe from attack.  We have created this contest to benchmark the latest computer vision and deep learning approaches to this problem. Can you crack the CAPTCHA? Can you improve the state of the art? Can you create lasting peace between cats and dogs?
Okay, well settle for the former. 
Acknowledgements
We extend our thanks to Microsoft Research for providing the data for this competition.",image data,playground,Create an algorithm to distinguish dogs from cats,CategorizationAccuracy,dogs-vs.-cats,categorization_accuracy
Flower Classification with TPUs,"Tensor Processing Units (TPUs) are Now Available on Kaggle
Tensor Processing Unit (TPU) quotas are now available on Kaggle, at no cost to you!
TPUs are powerful hardware accelerators specialized in deep learning tasks. They were developed (and first used) by Google to process large image databases, such as extracting all the text from Street View. This competition is designed for you to give TPUs a try.
The latest Tensorflow release (TF 2.1) was focused on TPUs and theyre now supported both through the Keras high-level API and at a lower level, in models using a custom training loop.
We cant wait to see how your solutions are accelerated by TPUs!
The Challenge
Its difficult to fathom just how vast and diverse our natural world is.
There are over 5,000 species of mammals, 10,000 species of birds, 30,000 species of fish  and astonishingly, over 400,000 different types of flowers.
In this competition, youre challenged to build a machine learning model that identifies the type of flowers in a dataset of images (for simplicity, were sticking to just over 100 types).
To get started with TPUs:
Read the TPU documentation one-pager
Then jump right into the Getting Started Notebook for this competition
Quick note: a TPU is a network-connected accelerator and requires a couple extra lines in your code. Flipping the TPU switch in your notebook will not, by itself, accelerate your code.
Have Questions?
Martin Grner, Google Developer Advocate and author of Tensorflow without a PhD will be actively engaged in the competition forum. If you have a question or need help troubleshooting, thats the best place to find help.",image data,playground,Use TPUs to classify 104 types of flowers,MacroFScore,flower-classification-with-tpus,f_score_multiclass
Generative Dog Images,"This competition is closed and no longer accepting submissions. The private leaderboard has been finalized as of 8/28/2019.
Important Warning: This competition has an experimental format and submission style (images as submission). Competitors must use generative methods to create their submission images and are not permitted to make submissions that include any images already classified as dogs or altered versions of such images.
To enforce and prevent cheating, we reserve the right to: (a) Visually inspect all participants submitted images, (b) review any submitted source code, (c) use these reviews to identify violators or determine winners, and (d) disqualify participants from the competition who are found in violation. This is also specified in the competitions rules
Use your training skills to create images, rather than identify them. Youll be using GANs, which are at the creative frontier of machine learning. You might think of GANs as robot artists in a senseable to create eerily lifelike images, and even digital worlds.
""You might not think that programmers are artists, but programming is an extremely creative profession. Its logic-based creativity.  -
John Romero
A generative adversarial network (GAN) is a class of machine learning system invented by Ian Goodfellow in 2014. Two neural networks compete with each other in a game. Given a training set, this technique learns to generate new data with the same statistics as the training set.
In this competition, youll be training generative models to create images of dogs. Only this time theres no ground truth data for you to predict. Here, youll submit the images and be scored based on how well those images are classified as dogs from pre-trained neural networks. Take these images, for example. Can you tell which are real vs. generated?
Trick question; they are all generated!
Why dogs? We chose dogs because, well, who doesnt love looking at photos of adorable pups? Moreover, dogs can be classified into many sub-categories (breed, color, size), making them ideal candidates for image generation.
Generative methods (in particular, GANs) are currently used in various places on Kaggle for data augmentation. Their potential is vast; they can learn to mimic any distribution of data across any domain: photographs, drawings, music, and prose. If successful, not only will you help advance the state of the art in generative image creation, but youll enable us to create more experiments across a variety of domains in the future.
This is a Kernels-only competition. Refer to Kernels Requirements for details.",image data,research,Experiment with creating puppy pics,PostProcessorKernel,generative-dog-images,custom_loss
"Quick, Draw! Doodle Recognition Challenge","""Quick, Draw!"" was released as an experimental game to educate the public in a playful way about how AI works. The game prompts users to draw an image depicting a certain category, such as banana, table, etc. The game generated more than 1B drawings, of which a subset was publicly released as the basis for this competitions training set. That subset contains 50M drawings encompassing 340 label categories.
Sounds fun, right? Heres the challenge: since the training data comes from the game itself, drawings can be incomplete or may not match the label. Youll need to build a recognizer that can effectively learn from this noisy data and perform well on a manually-labeled test set from a different distribution.
Your task is to build a better classifier for the existing Quick, Draw! dataset. By advancing models on this dataset, Kagglers can improve pattern recognition solutions more broadly. This will have an immediate impact on handwriting recognition and its robust applications in areas including OCR (Optical Character Recognition), ASR (Automatic Speech Recognition) & NLP (Natural Language Processing).",image data,featured,How accurately can you identify a doodle?,MAP@{K},"quick,-draw!-doodle-recognition-challenge",reconstruction
RANZCR CLiP - Catheter and Line Position Challenge,"Serious complications can occur as a result of malpositioned lines and tubes in patients. Doctors and nurses frequently use checklists for placement of lifesaving equipment to ensure they follow protocol in managing patients. Yet, these steps can be time consuming and are still prone to human error, especially in stressful situations when hospitals are at capacity.
Hospital patients can have catheters and lines inserted during the course of their admission and serious complications can arise if they are positioned incorrectly. Nasogastric tube malpositioning into the airways has been reported in up to 3% of cases, with up to 40% of these cases demonstrating complications [1-3]. Airway tube malposition in adult patients intubated outside the operating room is seen in up to 25% of cases [4,5]. The likelihood of complication is directly related to both the experience level and specialty of the proceduralist. Early recognition of malpositioned tubes is the key to preventing risky complications (even death), even more so now that millions of COVID-19 patients are in need of these tubes and lines.
The gold standard for the confirmation of line and tube positions are chest radiographs. However, a physician or radiologist must manually check these chest x-rays to verify that the lines and tubes are in the optimal position. Not only does this leave room for human error, but delays are also common as radiologists can be busy reporting other scans. Deep learning algorithms may be able to automatically detect malpositioned catheters and lines. Once alerted, clinicians can reposition or remove them to avoid life-threatening complications.
The Royal Australian and New Zealand College of Radiologists (RANZCR) is a not-for-profit professional organisation for clinical radiologists and radiation oncologists in Australia, New Zealand, and Singapore. The group is one of many medical organisations around the world (including the NHS) that recognizes malpositioned tubes and lines as preventable. RANZCR is helping design safety systems where such errors will be caught.
In this competition, youll detect the presence and position of catheters and lines on chest x-rays. Use machine learning to train and test your model on 40,000 images to categorize a tube that is poorly placed.
The dataset has been labelled with a set of definitions to ensure consistency with labelling. The normal category includes lines that were appropriately positioned and did not require repositioning. The borderline category includes lines that would ideally require some repositioning but would in most cases still function adequately in their current position. The abnormal category included lines that required immediate repositioning.
If successful, your efforts may help clinicians save lives. Earlier detection of malpositioned catheters and lines is even more important as COVID-19 cases continue to surge. Many hospitals are at capacity and more patients are in need of these tubes and lines. Quick feedback on catheter and line placement could help clinicians better treat these patients. Beyond COVID-19, detection of line and tube position will ALWAYS be a requirement in many ill hospital patients.
This is a Code Competition. Refer to Code Requirements for details.
Koopmann MC, Kudsk KA, Szotkowski MJ, Rees SM. A Team-Based Protocol and Electromagnetic Technology Eliminate Feeding Tube Placement Complications [Internet]. Vol. 253, Annals of Surgery. 2011. p. 297302. Available from: http://dx.doi.org/10.1097/sla.0b013e318208f550
Sorokin R, Gottlieb JE. Enhancing patient safety during feeding-tube insertion: a review of more than 2,000 insertions. JPEN J Parenter Enteral Nutr. 2006 Sep;30(5):4405.
Marderstein EL, Simmons RL, Ochoa JB. Patient safety: effect of institutional protocols on adverse events related to feeding tube placement in the critically ill. J Am Coll Surg. 2004 Jul;199(1):3947; discussion 4750.
Jemmett ME. Unrecognized Misplacement of Endotracheal Tubes in a Mixed Urban to Rural Emergency Medical Services Setting [Internet]. Vol. 10, Academic Emergency Medicine. 2003. p. 9615. Available from: http://dx.doi.org/10.1197/s1069-6563(03)00315-4
Lotano R, Gerber D, Aseron C, Santarelli R, Pratter M. Utility of postintubation chest radiographs in the intensive care unit. Crit Care. 2000 Jan 24;4(1):503.",image data,featured,Classify the presence and correct placement of tubes on chest x-rays to save lives,MCAUC,ranzcr-clip-catheter-and-line-position-challenge,auc_multiclass
The Nature Conservancy Fisheries Monitoring,"Nearly half of the world depends on seafood for their main source of protein. In the Western and Central Pacific, where 60% of the worlds tuna is caught, illegal, unreported, and unregulated fishing practices are threatening marine ecosystems, global seafood supplies and local livelihoods. The Nature Conservancy is working with local, regional and global partners to preserve this fishery for the future.
Currently, the Conservancy is looking to the future by using cameras to dramatically scale the monitoring of fishing activities to fill critical science and compliance monitoring data gaps. Although these electronic monitoring systems work well and are ready for wider deployment, the amount of raw data produced is cumbersome and expensive to process manually.
The Conservancy is inviting the Kaggle community to develop algorithms to automatically detect and classify species of tunas, sharks and more that fishing boats catch, which will accelerate the video review process. Faster review and more reliable data will enable countries to reallocate human capital to management and enforcement activities which will have a positive impact on conservation and our planet.
Machine learning has the ability to transform what we know about our oceans and how we manage them. You can be part of the solution.
Resources
You can learn more about this competition and The Nature Conservancy in the video below.",image data,featured,Can you detect and classify species of fish?,MulticlassLoss,the-nature-conservancy-fisheries-monitoring,multiclass_log_loss
Petals to the Metal - Flower Classification on TPU,"Learn how to use Tensor Processing Units (TPUs) on Kaggle
TPUs are powerful hardware accelerators specialized in deep learning tasks. They were developed (and first used) by Google to process large image databases, such as extracting all the text from Street View. This competition is designed for you to give TPUs a try.
TPU quotas are available on Kaggle at no cost to users.
Watch the video below to see how to get started! You can follow along with this notebook.
The Challenge
Its difficult to fathom just how vast and diverse our natural world is.
There are over 5,000 species of mammals, 10,000 species of birds, 30,000 species of fish  and astonishingly, over 400,000 different types of flowers.
In this competition, youre challenged to build a machine learning model that identifies the type of flowers in a dataset of images (for simplicity, were sticking to just over 100 types).
Recommended Tutorial
We highly recommend Ryan Holbrooks Tutorial that walks you through making your very first submission step by step.
Have Questions?
Kaggle Data Scientists will be actively monitoring the competition forum - your fellow data scientists and TPU users will be there too! If you have a question or need help troubleshooting, thats the best place to find help.
Learn More
Check out Kaggles Youtube playlist for more videos introducing TPUs.
Read the TPU documentation for more information and resources.
Many thanks to Martin Grner, Google Developer Advocate and author of Tensorflow without a PhD for his tireless work on the dataset, the notebooks, and the original competition that this Getting Started competition draws from.",image data,getting started,Getting Started with TPUs on Kaggle!,macrofscore,petals-to-the-metal-flower-classification-on-tpu,f_score_multiclass
Ultrasound Nerve Segmentation,"Even the bravest patient cringes at the mention of a surgical procedure. Surgery inevitably brings discomfort, and oftentimes involves significant post-surgical pain. Currently, patient pain is frequently managed through the use of narcotics that bring a bevy of unwanted side effects.
This competitions sponsor is working to improve pain management through the use of indwelling catheters that block or mitigate pain at the source. Pain management catheters reduce dependence on narcotics and speed up patient recovery.
Accurately identifying nerve structures in ultrasound images is a critical step in effectively inserting a patients pain management catheter. In this competition, Kagglers are challenged to build a model that can identify nerve structures in a dataset of ultrasound images of the neck. Doing so would improve catheter placement and contribute to a more pain free future. ",image data,featured,Identify nerve structures in ultrasound images of the neck,Dice,ultrasound-nerve-segmentation,segmentation
VietAI Advance Course - Retinal Disease Detection,"Overview
Color-filtered fundus images visualize the rear of an eye called retina (Figure 1). Fundus image provides doctors with a snapshot on the interior of the eye of patients. Based on this type of image, doctor will be able to read abnormalities present on the back of the eye, thus making diagnosis easier and more accurate. Many eye diseases can be found using fundus images, such as diabetic retinopathy, glaucoma, and macular degeneration.
Current available public datasets (EYEPACS, Messidor, etc.), although rich in quantity, only focus on diabetic retinopathy. However, more often than not, a patient can have two or more diseases concurrently. To overcome these 2 problems, we take other common diseases into consideration and create a multi-labeled dataset. The following will describe the dataset in details.
Problem description
The dataset includes 3,285 images from CTEH (3.210 abnormals and 75 normals) and 500 normal images from Messidor and EYEPACS dataset. The abnormalities include: opacity, diabetic retinopathy, glaucoma, macular edema, macular degeneration, and retinal vascular occlusion
In this assignment, we will use 3,435 images for training and predict diseases on 350 unlabeled images.
Acknowledgements
We thank AI Department  Cao Thang Eye Hospital (CTEH) for providing this dataset",image data,inclass,Assignment for VietAI Advance Course 2020 - Deep Learning in Vision,meanfscore,vietai-advance-course-retinal-disease-detection,f_score_multiclass
WiDS Datathon 2019,"In advance of the March 4, 2019 Global WiDS Conference, the Global WiDS team, the West Big Data Innovation Hub, and the WiDS Datathon Committee have been working with Planet and Figure Eight to bring a dataset of high-resolution satellite imagery to participants, building awareness about deforestation and oil palm plantations.
We invite you to build a team, hone your data science skills, and join us in this predictive analytics challenge focused on social impact. Keep reading to learn more about the datathon, the significance of oil palm, and how to get started.
UPDATE: The Sign Up and Merger Deadline is now Feb. 24 11:59PM UTC, but please note Kaggle has scheduled maintenance Feb. 22 5PM UTC - Feb. 23 5AM UTC.
Also, the official WiDS 2019 Datathon Participant Form and opportunity to be eligible for additional prizes is now up at http://bit.ly/WiDSdatathon2019form
Why oil palm?
Deforestation through oil palm plantation growth represents an agricultural trend with large economic and environmental impacts. From shampoo to donuts and ice cream, oil palm is present in many everyday productsbut many have never heard of it explicitly! Because oil palm grows only in tropical environments, the crops expansion has led to deforestation, increased carbon emissions, and biodiversity loss, while at the same time providing many valuable jobs.
With the economic livelihoods of millions and the ecosystems of the tropics at stake,
how might we work towards affordable, timely, and scalable ways to
address the expansion and management of oil palm throughout the world?
High-resolution satellite imagery is a global, regularly-updated, and accurate source of data. Coupled with computer vision algorithms, it presents a promising opportunity for automated mapping of oil palm plantations, an important step toward understanding global impact.
Who can participate
We invite anyone from those new to data science to veterans of the field to participate. For those who have never tried machine learning or worked with satellite data before, we will be releasing a series of guides to help you get started with the algorithms and dataset. Get the latest resources by following #WiDSDatathon on social media and visiting widsconference.org/datathon.
The WiDS Datathon aims to inspire women worldwide to learn more about data science, and to create a supportive environment for women to connect with others in their community who share their interests. Toward these ends, we open the datathon to individuals or teams of up to 4, at least half of each team must be women (individuals identifying as female). Participants can be students, faculty, government workers, members of NGOs, or industry members.
More details
The challenge is to create a model that predicts the presence of oil palm plantations in satellite imagery. Planet and Figure Eight have generously provided an annotated dataset of satellite images recently taken by Planet satellites. The dataset images are 3-meter spatial resolution, and each is labeled with whether an oil palm plantation appears in the image (0 for no plantation, 1 for any presence of a plantation).
The datathon task is to train a model that takes as input a satellite image and outputs a prediction of how likely it is that the image contains an oil palm plantation. Labeled training and test datasets are provided for model development; you will then upload your predictions for an unlabeled test set to Kaggle and these predictions will be used to determine the public leaderboard rankings, and the final winners of the competition.
Data analysis can be done using your preferred tools. The winners will be determined by the leaderboard on the Kaggle platform at the time the contest closes February 27.
For more details and answers to frequently asked questions, please visit our FAQ page
Acknowledgements
The WiDS Datathon 2019 is a collaboration led by the Global WiDS team at Stanford, the West Big Data Innovation Hub, and the WiDS Datathon Committee. Special thanks to data providers Planet and Figure Eight, as well as our growing community of sponsors and supporters.",image data,inclass,Join the Women in Data Science (WiDS) Datathon 2019,auc,wids-datathon-2019,auc
iWildCam 2019 - FGVC6,"Camera Traps (or Wild Cams) enable the automatic collection of large quantities of image data. Biologists all over the world use camera traps to monitor biodiversity and population density of animal species. We have recently been making strides towards automating the species classification challenge in camera traps, but as we try to expand the scope of these models from specific regions where we have collected training data to nearby areas we are faced with an interesting probem: how do you classify a species in a new region that you may not have seen in previous training data?
In order to tackle this problem, we have prepared a challenge where the training data and test data are from different regions, namely The American Southwest and the American Northwest. The species seen in each region overlap, but are not identical, and the challenge is to classify the test species correctly. To this end, we will allow training on our American Southwest data (from CaltechCameraTraps), on iNaturalist 2017/2018 data, and on simulated data generated from Microsoft AirSim. We have provided a taxonomy file mapping our classes into the iNat taxonomy.
This is an FGVCx competition as part of the FGVC6 workshop at CVPR 2019, and is sponsored by Microsoft AI for Earth. There is a github page for the competition here. Please open an issue if you have questions or problems with the dataset.
If you use this dataset in publication, please cite:
@article{beery2019iwildcam,
 title={The iWildCam 2019 Challenge Dataset},
 author={Beery, Sara and Morris, Dan and Perona, Pietro},
 journal={arXiv preprint arXiv:1907.07617},
 year={2019}
}

Kaggle is excited to partner with research groups to push forward the frontier of machine learning. Research competitions make use of Kaggles platform and experience, but are largely organized by the research groups data science team. Any questions or concerns regarding the competition data, quality, or topic will be addressed by them.",image data,playground,Categorize animals in the wild,MacroFScore,iwildcam-2019-fgvc6,f_score_multiclass
Hackathon Auto_matic,"Virtual Hackathon
Participate in virtual hackathon for scholars of Secure and Private AI Scholarship Challenge from Facebook conducted by #sghackathonorgnizrs. Come join us for a fun filled weekend of coding and competing against each other.
When is it?
Hackathon starts => Saturday 00:01am GMT to Monday 11:59am GMT
Coding Time => Saturday 00:01am GMT to Sunday 11:59pm GMT .
Commiting Kernel => Sunday 00:01am GMT to Monday 11:59am GMT
classification
How to participate?

Use this form to sign up. You can participate alone or as part of a team of up to 4 individuals. Only 1 member of the team needs to fill the form. http://bit.ly/hackathon-signup. Please join the #sg_hackathon-orgnizrs channel to get the announcements and ask questions.
5. When will results be announced?
On Wednesday
Acknowledgements
We thank Udacity and Facebook for this opportunity
For more FAQs, please go to our our github page",image data,inclass,Cars Dataset,categorizationaccuracy,hackathon-auto_matic,categorization_accuracy
iWildCam 2020 - FGVC7,"Camera Traps (or Wild Cams) enable the automatic collection of large quantities of image data. Biologists all over the world use camera traps to monitor biodiversity and population density of animal species. We have recently been making strides towards automatic species classification in camera trap images. However, as we try to expand the scope of these models we are faced with an interesting problem: how do we train models that perform well on new (unseen during training) camera trap locations? Can we leverage data from other modalities, such as citizen science data and remote sensing data?
In order to tackle this problem, we have prepared a challenge where the training data and test data are from different cameras spread across the globe. The set of species seen in each camera overlap, but are not identical. The challenge is to classify species in the test cameras correctly. To explore multimodal solutions, we allow competitors to train on the following data: (i) our camera trap training set (data provided by WCS), (ii) iNaturalist 2017-2019 data, and (iii) multispectral imagery (from Landsat 8) for each of the camera trap locations. On the competition GitHub page we provide the multispectral data, a taxonomy file mapping our classes into the iNat taxonomy, a subset of iNat data mapped into our class set, and a camera trap detection model (the MegaDetector) along with the corresponding detections.
If you use this dataset in publication, please cite:
@article{beery2020iwildcam,
    title={The iWildCam 2020 Competition Dataset},
    author={Beery, Sara and Cole, Elijah and Gjoka, Arvi},
    journal={arXiv preprint arXiv:2004.10340},
    year={2020}
}
This is an FGVCx competition as part of the FGVC7 workshop at CVPR 2020, and is sponsored by Microsoft AI for Earth and Wildlife Insights. There is a GitHub page for the competition here. Please open an issue if you have questions or problems with the dataset.
You can find the iWildCam 2018 Competition here, and the iWildCam 2019 Competition here.
Kaggle is excited to partner with research groups to push forward the frontier of machine learning. Research competitions make use of Kaggles platform and experience, but are largely organized by the research groups data science team. Any questions or concerns regarding the competition data, quality, or topic will be addressed by them.",image data,research,Categorize animals in the wild,CategorizationAccuracy,iwildcam-2020-fgvc7,categorization_accuracy
Understanding Clouds from Satellite Images,"Climate change has been at the top of our minds and on the forefront of important political decision-making for many years. We hope you can use this competitions dataset to help demystify an important climatic variable. Scientists, like those at Max Planck Institute for Meteorology, are leading the charge with new research on the worlds ever-changing atmosphere and they need your help to better understand the clouds.
Shallow clouds play a huge role in determining the Earths climate. Theyre also difficult to understand and to represent in climate models. By classifying different types of cloud organization, researchers at Max Planck hope to improve our physical understanding of these clouds, which in turn will help us build better climate models.
There are many ways in which clouds can organize, but the boundaries between different forms of organization are murky. This makes it challenging to build traditional rule-based algorithms to separate cloud features. The human eye, however, is really good at detecting featuressuch as clouds that resemble flowers.
In this challenge, you will build a model to classify cloud organization patterns from satellite images. If successful, youll help scientists to better understand how clouds will shape our future climate. This research will guide the development of next-generation models which could reduce uncertainties in climate projections.
Help us remove the haze from climate models and bring clarity to cloud identification.
For more information on the scientific background and how the labels were created see the following paper.",image data,research,Can you classify cloud structures from satellites? ,Dice,understanding-clouds-from-satellite-images,segmentation
VinBigData Chest X-ray Abnormalities Detection,"When you have a broken arm, radiologists help save the dayand the bone. These doctors diagnose and treat medical conditions using imaging techniques like CT and PET scans, MRIs, and, of course, X-rays. Yet, as it happens when working with such a wide variety of medical tools, radiologists face many daily challenges, perhaps the most difficult being the chest radiograph. The interpretation of chest X-rays can lead to medical misdiagnosis, even for the best practicing doctor. Computer-aided detection and diagnosis systems (CADe/CADx) would help reduce the pressure on doctors at metropolitan hospitals and improve diagnostic quality in rural areas.
Existing methods of interpreting chest X-ray images classify them into a list of findings. There is currently no specification of their locations on the image which sometimes leads to inexplicable results. A solution for localizing findings on chest X-ray images is needed for providing doctors with more meaningful diagnostic assistance.
Established in August 2018 and funded by the Vingroup JSC, the Vingroup Big Data Institute (VinBigData) aims to promote fundamental research and investigate novel and highly-applicable technologies. The Institute focuses on key fields of data science and artificial intelligence: computational biomedicine, natural language processing, computer vision, and medical image processing. The medical imaging team at VinBigData conducts research in collecting, processing, analyzing, and understanding medical data. Theyre working to build large-scale and high-precision medical imaging solutions based on the latest advancements in artificial intelligence to facilitate effective clinical workflows.
In this competition, youll automatically localize and classify 14 types of thoracic abnormalities from chest radiographs. Youll work with a dataset consisting of 18,000 scans that have been annotated by experienced radiologists. You can train your model with 15,000 independently-labeled images and will be evaluated on a test set of 3,000 images. These annotations were collected via VinBigDatas web-based platform, VinLab. Details on building the dataset can be found in our recent paper VinDr-CXR: An open dataset of chest X-rays with radiologists annotations.
If successful, youll help build what could be a valuable second opinion for radiologists. An automated system that could accurately identify and localize findings on chest radiographs would relieve the stress of busy doctors while also providing patients with a more accurate diagnosis.
Acknowledgments
Challenge Organizing Team
Ha Q. Nguyen, PhD - Vingroup Big Data Institute
Hieu H. Pham, PhD - Vingroup Big Data Institute
Nhan T. Nguyen, MSc - Vingroup Big Data Institute
Dung B. Nguyen, BSc - Vingroup Big Data Institute
Minh Dao, PhD - Vingroup Big Data Institute
Van Vu, PhD - Vingroup Big Data Institute
Khanh Lam, MD, PhD - Hospital 108
Linh T. Le, MD, PhD - Hanoi Medical University Hospital
Data Contributors
The dataset used in this competition was created by assembling de-identified Chest X-ray studies provided by two hospitals in Vietnam: the Hospital 108 and the Hanoi Medical University Hospital.",image data,featured,Automatically localize and classify thoracic abnormalities from chest radiographs,OpenImagesObjectDetectionAP,vinbigdata-chest-x-ray-abnormalities-detection,detection
Global Wheat Detection,"Open up your pantry and youre likely to find several wheat products. Indeed, your morning toast or cereal may rely upon this common grain. Its popularity as a food and crop makes wheat widely studied. To get large and accurate data about wheat fields worldwide, plant scientists use image detection of ""wheat heads""spikes atop the plant containing grain. These images are used to estimate the density and size of wheat heads in different varieties. Farmers can use the data to assess health and maturity when making management decisions in their fields.
However, accurate wheat head detection in outdoor field images can be visually challenging. There is often overlap of dense wheat plants, and the wind can blur the photographs. Both make it difficult to identify single heads. Additionally, appearances vary due to maturity, color, genotype, and head orientation. Finally, because wheat is grown worldwide, different varieties, planting densities, patterns, and field conditions must be considered. Models developed for wheat phenotyping need to generalize between different growing environments. Current detection methods involve one- and two-stage detectors (Yolo-V3 and Faster-RCNN), but even when trained with a large dataset, a bias to the training region remains.
The Global Wheat Head Dataset is led by nine research institutes from seven countries: the University of Tokyo, Institut national de recherche pour lagriculture, lalimentation et lenvironnement, Arvalis, ETHZ, University of Saskatchewan, University of Queensland, Nanjing Agricultural University, and Rothamsted Research. These institutions are joined by many in their pursuit of accurate wheat head detection, including the Global Institute for Food Security, DigitAg, Kubota, and Hiphen.
In this competition, youll detect wheat heads from outdoor images of wheat plants, including wheat datasets from around the globe. Using worldwide data, you will focus on a generalized solution to estimate the number and size of wheat heads. To better gauge the performance for unseen genotypes, environments, and observational conditions, the training dataset covers multiple regions. You will use more than 3,000 images from Europe (France, UK, Switzerland) and North America (Canada). The test data includes about 1,000 images from Australia, Japan, and China.
Wheat is a staple across the globe, which is why this competition must account for different growing conditions. Models developed for wheat phenotyping need to be able to generalize between environments. If successful, researchers can accurately estimate the density and size of wheat heads in different varieties. With improved detection farmers can better assess their crops, ultimately bringing cereal, toast, and other favorite dishes to your table.
This is a Code Competition. Refer to Code Requirements for details.",image data,research,Can you help identify wheat heads using image analysis?,custom metric,global-wheat-detection,detection
Human Protein Atlas - Single Cell Classification,"There are billions of humans on this earth, and each of us is made up of trillions of cells. Just like every individual is unique, even genetically identical twins, scientists observe differences between the genetically identical cells in our bodies.
Differences in the location of proteins can give rise to such cellular heterogeneity. Proteins play essential roles in virtually all cellular processes. Often, many different proteins come together at a specific location to perform a task, and the exact outcome of this task depends on which proteins are present. As you can imagine, different subcellular distributions of one protein can give rise to great functional heterogeneity between cells. Finding such differences, and figuring out how and why they occur, is important for understanding how cells function, how diseases develop, and ultimately how to develop better treatments for those diseases.
To see more, start with less. That may seem counterintuitive, but the study of a single cell enables the discovery of mechanisms too difficult to see with multi-cell research. The importance of studying single cells is reflected in the ongoing revolution in biology centered around technologies for single cell analysis. Microscopy offers an opportunity to study differences in protein localizations within a population of cells. Current machine learning models for classifying protein localization patterns in microscope images gives a summary of the entire population of cells. However, the single-cell revolution in biology demands models that can precisely classify patterns in each individual cell in the image.
The Human Protein Atlas is an initiative based in Sweden that is aimed at mapping proteins in all human cells, tissues, and organs. The data in the Human Protein Atlas database is freely accessible to scientists all around the world that allows them to explore the cellular makeup of the human body. Solving the single-cell image classification challenge will help us characterize single-cell heterogeneity in our large collection of images by generating more accurate annotations of the subcellular localizations for thousands of human proteins in individual cells. Thanks to you, we will be able to more accurately model the spatial organization of the human cell and provide new open-access cellular data to the scientific community, which may accelerate our growing understanding of how human cells functions and how diseases develop.
This is a weakly supervised multi-label classification problem and a code competition. Given images of cells from our microscopes and labels of protein location assigned together for all cells in the image, Kagglers will develop models capable of segmenting and classifying each individual cell with precise labels. If successful, youll contribute to the revolution of single-cell biology!
The scientific journal Nature Methods is interested in considering a paper discussing the outcome and approaches of the challenge. The Human Protein Atlas team, led by Professor Emma Lundberg, would like to invite top performing teams to join as co-authors in writing this paper. Please follow the discussion forum for more details on how you can help.
This is a Code Competition. Refer to Code Requirements for details.",image data,featured,Find individual human cell differences in microscope images,OpenImagesObjDetectionSegmentationAP,human-protein-atlas-single-cell-classification,segmentation
HuBMAP - Hacking the Kidney,"Our best estimates show there are over 7 billion people on the planet and 300 billion stars in the Milky Way galaxy. By comparison, the adult human body contains 37 trillion cells. To determine the function and relationship among these cells is a monumental undertaking. Many areas of human health would be impacted if we better understand cellular activity. A problem with this much data is a great match for the Kaggle community.
Just as the Human Genome Project mapped the entirety of human DNA, the Human BioMolecular Atlas Program (HuBMAP) is a major endeavor. Sponsored by the National Institutes of Health (NIH), HuBMAP is working to catalyze the development of a framework for mapping the human body at a level of glomeruli functional tissue units for the first time in history. Hoping to become one of the worlds largest collaborative biological projects, HuBMAP aims to be an open map of the human body at the cellular level.
This competition, Hacking the Kidney,"" starts by mapping the human kidney at single cell resolution.
Your challenge is to detect functional tissue units (FTUs) across different tissue preparation pipelines. An FTU is defined as a three-dimensional block of cells centered around a capillary, such that each cell in this block is within diffusion distance from any other cell in the same block (de Bono, 2013). The goal of this competition is the implementation of a successful and robust glomeruli FTU detector.
You will also have the opportunity to present your findings to a panel of judges for additional consideration. Successful submissions will construct the tools, resources, and cell atlases needed to determine how the relationships between cells can affect the health of an individual.
Advancements in HuBMAP will accelerate the worlds understanding of the relationships between cell and tissue organization and function and human health. These datasets and insights can be used by researchers in cell and tissue anatomy, pharmaceutical companies to develop therapies, or even parents to show their children the magnitude of the human body.
This is a Code Competition. Refer to Code Requirements for details.",image data,research,Identify glomeruli in human kidney tissue images,Dice,hubmap-hacking-the-kidney,segmentation
Human Protein Atlas Image Classification,"In this competition, Kagglers will develop models capable of classifying mixed patterns of proteins in microscope images. The Human Protein Atlas will use these models to build a tool integrated with their smart-microscopy system to identify a proteins location(s) from a high-throughput image.
Proteins are the doers in the human cell, executing many functions that together enable life. Historically, classification of proteins has been limited to single patterns in one or a few cell types, but in order to fully understand the complexity of the human cell, models must classify mixed patterns across a range of different human cells.
Images visualizing proteins in cells are commonly used for biomedical research, and these cells could hold the key for the next breakthrough in medicine. However, thanks to advances in high-throughput microscopy, these images are generated at a far greater pace than what can be manually evaluated. Therefore, the need is greater than ever for automating biomedical image analysis to accelerate the understanding of human cells and disease.
  Nature Methods has indicated interest in considering a paper discussing the outcome and approaches of the challenge. The Human Protein Atlas team would like to invite top performing teams to join as co-authors in the writing of this paper.
Top performing teams will also be eligible to compete for the special prize. Additional information for both the special prize and co-authoring for Nature Methods will become available through the Discussion posts once the main competition is complete.
  Acknowledgements
The Human Protein Atlas is a Sweden-based initiative aimed at mapping all human proteins in cells, tissues and organs. All the data in the knowledge resource is open access to allow anyone to pursue exploration of the human proteome. In a recent publication, the Human Protein Atlas team has demonstrated the promise of both citizen science and artificial intelligence approaches in describing the location of human proteins in images, however current results are yet to approach expert-level annotations (Sullivan et al, Nature Biotechnology, Oct 2018).",image data,featured,Classify subcellular protein patterns in human cells,MacroFScore,human-protein-atlas-image-classification,f_score_multiclass
iFood - 2019 at FGVC6,"What did you eat today? Wondering if you are eating a healthy diet? Automatic food identification can assist towards food intake monitoring to maintain a healthy diet. Food classification is a challenging problem due to the large number of food categories, high visual similarity between different food categories, as well as the lack of datasets that are large enough for training deep models. In this competition, we extend our last years dataset to 251 fine-grained (prepared) food categories with 118,475 training images collected from the web. We provide human verified labels for both the validation set of 11,994 images and the test set of 28,377 images. The goal is to build a model to predict the fine-grained food-category label given an image.
The main challenges are:
Fine-grained Classes: The classes are fine-grained and visually similar. For example, the dataset has 15 different types of cakes, and 10 different types of pastas.
Noisy Data: Since the training images are crawled from the web, they often include images of raw ingredients or processed and packaged food items. This is referred to as cross-domain noise. Further, due to the fine-grained nature of food-categories, a training image may either be incorrectly labeled into a visually similar class or be annotated with with a single label despite having multiple food items.
This competition is part of the fine-grained visual-categorization workshop (FGVC6 workshop) at CVPR 2019 . There is a Github page for the competition here. For any queries you can start a discussion topic here or email us at ifood2019@gmail.com.
Participants who make a submission that beats the sample submission can fill out this form to receive $150 in Google Cloud credits.
We would like to thank SRI International and Google for support in data collection and labeling. The challenge is sponsored by SRI International.",image data,inclass,Fine-grained classification of food images,meanbesterroratk,ifood-2019-at-fgvc6,reconstruction
iNaturalist 2019 at FGVC6,"As part of the FGVC6 workshop at CVPR 2019 we are conducting the iNat Challenge 2019 large scale species classification competition, sponsored by Microsoft. It is estimated that the natural world contains several million species of plants and animals. Without expert knowledge, many of these species are extremely difficult to accurately classify due to their visual similarity. The goal of this competition is to push the state of the art in automatic image classification for real world data that features a large number of fine-grained categories.
Previous versions of the challenge have focused on classifying large numbers of species. This year features a smaller number of highly similar categories captured in a wide variety of situations, from all over the world. In total, the iNat Challenge 2019 dataset contains 1,010 species, with a combined training and validation set of 268,243 images that have been collected and verified by multiple users from iNaturalist.
Teams with top submissions, at the discretion of the workshop organizers, will be invited to present their work at the FGVC6 workshop. Participants who make a submission that beats the sample submission can fill out this form to receive $150 in Google Cloud credits.

Kaggle is excited to partner with research groups to push forward the frontier of machine learning. Research competitions make use of Kaggles platform and experience, but are largely organized by the research groups data science team. Any questions or concerns regarding the competition data, quality, or topic will be addressed by them.",image data,research,Fine-grained classification spanning a thousand species,MeanBestErrorAtK,inaturalist-2019-at-fgvc6,reconstruction
Painter by Numbers,"With an original Picasso carrying a 106 million dollar price tag, identifying an authentic work of art from a forgery is a high-stakes industry. While algorithms have gotten good at telling us if a still life is of a basket of apples or a sunflower bouquet, they arent yet able to tell us with certainty if both paintings are by van Gogh.  
In this playground competition, were challenging Kagglers to examine pairs of paintings and determine if they are by the same artist. This is an excellent opportunity to improve your computer vision skills and engage with a unique dataset of art. From the movement of brushstrokes to the use of light and dark, successful algorithms will likely incorporate many aspects of a painters unique style. 
Resources
neural algorithm
How Do We See Art: An Eye-Tracker Study
Acknowledgments
Many of the images in this dataset were obtained from wikiart.org. Additional paintings were provided by artists whose contributions will be acknowledged at the close of the competition.
This playground competition and its datasets were prepared by Small Yellow Duck (Kiri Nichol). This includes the design of the pairwise-evaluation scheme.",image data,playground,Does every painter leave a fingerprint? ,AUC,painter-by-numbers,auc
Cleaned vs Dirty,"Hi! It is boring to wash the dishes. Luckily, half of them are already clean. Train a classifier to determine clean ones to save time for the new machine learning course ;)
It is a few shot learning competition. We have a dataset of 20 clean and 20 dirty plates in train and hundreds of plates in test. Good luck!",image data,inclass,Classify if a plate is cleaned or dirty?,categorizationaccuracy,cleaned-vs-dirty,categorization_accuracy
Peking University/Baidu - Autonomous Driving,"Who do you think hates traffic more - humans or self-driving cars? The position of nearby automobiles is a key question for autonomous vehicles  and its at the heart of our newest challenge.
Self-driving cars have come a long way in recent years, but theyre still not flawless. Consumers and lawmakers remain wary of adoption, in part because of doubts about vehicles ability to accurately perceive objects in traffic.
Baidus Robotics and Autonomous Driving Lab (RAL), along with Peking University, hopes to close the gap once and for all with this challenge. Theyre providing Kagglers with more than 60,000 labeled 3D car instances from 5,277 real-world images, based on industry-grade CAD car models.
Your challenge: develop an algorithm to estimate the absolute pose of vehicles (6 degrees of freedom) from a single image in a real-world traffic environment.
Succeed and youll help improve computer vision. That, in turn, will bring autonomous vehicles a big step closer to widespread adoption, so they can help reduce the environmental impact of our growing societies.
Please cite the following paper when using the dataset:
ApolloCar3D: A Large 3D Car Instance Understanding Benchmark for Autonomous Driving
@inproceedings{song2019apollocar3d,
title={Apollocar3d: A large 3d car instance understanding benchmark for autonomous driving},
author={Song, Xibin and Wang, Peng and Zhou, Dingfu and Zhu, Rui and Guan, Chenye and Dai, Yuchao and Su, Hao and Li, Hongdong and Yang, Ruigang},
booktitle={Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition},
pages={5452--5462},
year={2019}
}",image data,featured,Can you predict vehicle angle in different settings?,PKUAutoDrivingAP,peking-university/baidu-autonomous-driving,detection
Open Images 2019 - Instance Segmentation,"Introduction
Computer vision has advanced considerably but is still challenged in matching the precision of human perception.
Open Images is a collaborative release of ~9 million images annotated with image-level labels, object bounding boxes, object segmentation masks, and visual relationships. This uniquely large and diverse dataset is designed to spur state of the art advances in analyzing and understanding images.
This years Open Images V5 release enabled the second Open Images Challenge to include the following 3 tracks:
Object detection track for detecting bounding boxes around object instances, relaunched from 2018.
Visual relationship detection track for detecting pairs of objects in particular relations, also relaunched from 2018.
Instance segmentation track for segmenting masks of objects in images, brand new for 2019.
Google AI hopes that having a single dataset with unified annotations for image classification, object detection, visual relationship detection, and instance segmentation will stimulate progress towards genuine scene understanding.
Instance Segmentation Track
In this track of the Challenge, you are asked to provide segmentation masks of objects.
This tracks training set represents 2.1M segmentation masks for object instances in 300 categories; with a validation set containing an additional 23k masks. The train set masks were produced by our state-of-the-art interactive segmentation process, where professional human annotators iteratively correct the output of a segmentation neural network. The validation and test set masks have been annotated manually with a strong focus on quality.
Example train set annotations. Left: Wuxi science park, 1995 by Gary Stevens. Right: Cat Cafe Shinjuku calico by Ari Helminen. Both images used under CC BY 2.0 license.

The results of this Challenge will be presented at a workshop at the International Conference on Computer Vision.
We are excited to partner with Open Images for this second year of competitions, including this brand new track!",image data,research,Outline segmentation masks of objects in images,OpenImagesObjDetectionSegmentationAP,open-images-2019-instance-segmentation,segmentation
Leaf Classification,"There are estimated to be nearly half a million species of plant in the world. Classification of species has been historically problematic and often results in duplicate identifications. Automating plant recognition might have many applications, including:
The objective of this playground competition is to use binary leaf images and extracted features, including shape, margin & texture, to accurately identify 99 species of plants. Leaves, due to their volume, prevalence, and unique characteristics, are an effective means of differentiating plant species. They also provide a fun introduction to applying techniques that involve image-based features.
As a first step, try building a classifier that uses the provided pre-extracted features. Next, try creating a set of your own features. Finally, examine the errors youre making and see what you can do to improve.
Acknowledgments
Kaggle is hosting this competition for the data science community to use for fun and education. This dataset originates from leaf images collected by  
James Cope, Thibaut Beghin, Paolo Remagnino, & Sarah Barman of the Royal Botanic Gardens, Kew, UK.
Charles Mallah, James Cope, James Orwell. Plant Leaf Classification Using Probabilistic Integration of Shape, Texture and Margin Features. Signal Processing, Pattern Recognition and Applications, in press. 2013.
We thank the UCI machine learning repository for hosting the dataset.",image data,playground,Can you see the random forest for the leaves?,MulticlassLoss,leaf-classification,multiclass_log_loss
ods_class_cs231n,CIFAR 10 playground.,image data,inclass,CIFAR10 playground,multiclassloss,ods_class_cs231n,multiclass_log_loss
Open Images 2019 - Object Detection,"Introduction
Computer vision has advanced considerably but is still challenged in matching the precision of human perception.
Open Images is a collaborative release of ~9 million images annotated with image-level labels, object bounding boxes, object segmentation masks, and visual relationships. This uniquely large and diverse dataset is designed to spur state of the art advances in analyzing and understanding images.
This years Open Images V5 release enabled the second Open Images Challenge to include the following 3 tracks:
Object detection track for detecting bounding boxes around object instances, relaunched from 2018.
Visual relationship detection track for detecting pairs of objects in particular relations, also relaunched from 2018.
Instance segmentation track for segmenting masks of objects in images, brand new for 2019.
Google AI hopes that having a single dataset with unified annotations for image classification, object detection, visual relationship detection, and instance segmentation will stimulate progress towards genuine scene understanding.
Object Detection Track
In this track of the Challenge, you are asked to predict a tight bounding box around object instances.
The training set contains 12.2M bounding-boxes across 500 categories on 1.7M images. The boxes have been largely manually drawn by professional annotators to ensure accuracy and consistency. The images are very diverse and often contain complex scenes with several objects (7 per image on average).

Example annotations. Left: Mark Paul Gosselaar plays the guitar by Rhys A. Right: the house by anita kluska. Both images used under CC BY 2.0 license.
Please refer to the Open Images 2019 Challenge page for additional details. The challenge contains a total of 3 tracks, which are linked above in the introduction. You are invited to explore and enter as many tracks as interest you.
The results of this Challenge will be presented at a workshop at the International Conference on Computer Vision.
We are excited to partner with Open Images for this second year of competitions. See link here for last years Object Detection competition.",image data,research,Detect objects in varied and complex images,OpenImagesObjectDetectionAP,open-images-2019-object-detection,detection
OSIC Pulmonary Fibrosis Progression,"Imagine one day, your breathing became consistently labored and shallow. Months later you were finally diagnosed with pulmonary fibrosis, a disorder with no known cause and no known cure, created by scarring of the lungs. If that happened to you, you would want to know your prognosis. Thats where a troubling disease becomes frightening for the patient: outcomes can range from long-term stability to rapid deterioration, but doctors arent easily able to tell where an individual may fall on that spectrum. Your help, and data science, may be able to aid in this prediction, which would dramatically help both patients and clinicians.
Current methods make fibrotic lung diseases difficult to treat, even with access to a chest CT scan. In addition, the wide range of varied prognoses create issues organizing clinical trials. Finally, patients suffer extreme anxietyin addition to fibrosis-related symptomsfrom the diseases opaque path of progression.
Open Source Imaging Consortium (OSIC) is a not-for-profit, co-operative effort between academia, industry and philanthropy. The group enables rapid advances in the fight against Idiopathic Pulmonary Fibrosis (IPF), fibrosing interstitial lung diseases (ILDs), and other respiratory diseases, including emphysematous conditions. Its mission is to bring together radiologists, clinicians and computational scientists from around the world to improve imaging-based treatments.
In this competition, youll predict a patients severity of decline in lung function based on a CT scan of their lungs. Youll determine lung function based on output from a spirometer, which measures the volume of air inhaled and exhaled. The challenge is to use machine learning techniques to make a prediction with the image, metadata, and baseline FVC as input.
If successful, patients and their families would better understand their prognosis when they are first diagnosed with this incurable lung disease. Improved severity detection would also positively impact treatment trial design and accelerate the clinical development of novel treatments.

This is a Code Competition. Refer to Code Requirements for details.",image data,featured,Predict lung function decline,LaplaceLogLikelihood,osic-pulmonary-fibrosis-progression,significance
Planet: Understanding the Amazon from Space,"Every minute, the world loses an area of forest the size of 48 football fields. And deforestation in the Amazon Basin accounts for the largest share, contributing to reduced biodiversity, habitat loss, climate change, and other devastating effects. But better data about the location of deforestation and human encroachment on forests can help governments and local stakeholders respond more quickly and effectively.
Planet, designer and builder of the worlds largest constellation of Earth-imaging satellites, will soon be collecting daily imagery of the entire land surface of the earth at 3-5 meter resolution. While considerable research has been devoted to tracking changes in forests, it typically depends on coarse-resolution imagery from Landsat (30 meter pixels) or MODIS (250 meter pixels). This limits its effectiveness in areas where small-scale deforestation or forest degradation dominate.
Furthermore, these existing methods generally cannot differentiate between human causes of forest loss and natural causes. Higher resolution imagery has already been shown to be exceptionally good at this, but robust methods have not yet been developed for Planet imagery.
In this competition, Planet and its Brazilian partner SCCON are challenging Kagglers to label satellite image chips with atmospheric conditions and various classes of land cover/land use. Resulting algorithms will help the global community better understand where, how, and why deforestation happens all over the world - and ultimately how to respond.
To dig into/explore more Planet data, sign up for a free account.
And if youre interested in building applications on Planet data, check out our Application Developer Program.
Getting Started
Review the data page, which includes detailed information about the labels and the labeling process.
Download a subsample of the data to get familiar with how it looks.
Explore the subsample on Kernels. Weve created a notebook for you to get started.",image data,featured,Use satellite data to track the human footprint in the Amazon rainforest,MeanFScoreBeta,planet:-understanding-the-amazon-from-space,f_score_multiclass
Plant Pathology 2020 - FGVC7,"Problem Statement
Misdiagnosis of the many diseases impacting agricultural crops can lead to misuse of chemicals leading to the emergence of resistant pathogen strains, increased input costs, and more outbreaks with significant economic loss and environmental impacts. Current disease diagnosis based on human scouting is time-consuming and expensive, and although computer-vision based models have the promise to increase efficiency, the great variance in symptoms due to age of infected tissues, genetic variations, and light conditions within trees decreases the accuracy of detection.
Specific Objectives
Objectives of Plant Pathology Challenge are to train a model using images of training dataset to 1) Accurately classify a given image from testing dataset into different diseased category or a healthy leaf; 2) Accurately distinguish between many diseases, sometimes more than one on a single leaf; 3) Deal with rare classes and novel symptoms; 4) Address depth perceptionangle, light, shade, physiological age of the leaf; and 5) Incorporate expert knowledge in identification, annotation, quantification, and guiding computer vision to search for relevant features during learning.
Resources
Details and background information on the dataset and Kaggle competition Plant Pathology 2020 Challenge were published. If you use the dataset for your project, please cite the following peer-reviewed research article
Thapa, Ranjita; Zhang, Kai; Snavely, Noah; Belongie, Serge; Khan, Awais. The Plant Pathology Challenge 2020 data set to classify foliar disease of apples. Applications in Plant Sciences, 8 (9), 2020.
Acknowledgments
We acknowledge financial support from Cornell Initiative for Digital Agriculture (CIDA) and special thanks to Zach Guillian for help with data collection.
Kaggle is excited to partner with research groups to push forward the frontier of machine learning. Research competitions make use of Kaggles platform and experience, but are largely organized by the research groups data science team. Any questions or concerns regarding the competition data, quality, or topic will be addressed by them.",image data,research,Identify the category of foliar diseases in apple trees,MCAUC,plant-pathology-2020-fgvc7,auc_multiclass
Plant Seedlings Classification,"Can you differentiate a weed from a crop seedling?
The ability to do so effectively can mean better crop yields and better stewardship of the environment.
The Aarhus University Signal Processing group, in collaboration with University of Southern Denmark, has recently released a dataset containing images of approximately 960 unique plants belonging to 12 species at several growth stages.
Were hosting this dataset as a Kaggle competition in order to give it wider exposure, to give the community an opportunity to experiment with different image recognition techniques, as well to provide a place to cross-pollenate ideas.
Acknowledgments
We extend our appreciation to the Aarhus University Department of Engineering Signal Processing Group for hosting the original data.
Citation
A Public Image Database for Benchmark of Plant Seedling Classification Algorithms",image data,playground,Determine the species of a seedling from an image,MeanFScore,plant-seedlings-classification,f_score_multiclass
SIIM-ACR Pneumothorax Segmentation,"Imagine suddenly gasping for air, helplessly breathless for no apparent reason. Could it be a collapsed lung? In the future, your entry in this competition could predict the answer.
Pneumothorax can be caused by a blunt chest injury, damage from underlying lung disease, or most horrifyingit may occur for no obvious reason at all. On some occasions, a collapsed lung can be a life-threatening event.
Pneumothorax is usually diagnosed by a radiologist on a chest x-ray, and can sometimes be very difficult to confirm. An accurate AI algorithm to detect pneumothorax would be useful in a lot of clinical scenarios. AI could be used to triage chest radiographs for priority interpretation, or to provide a more confident diagnosis for non-radiologists.
The Society for Imaging Informatics in Medicine (SIIM) is the leading healthcare organization for those interested in the current and future use of informatics in medical imaging. Their mission is to advance medical imaging informatics across the enterprise through education, research, and innovation in a multi-disciplinary community. Today, they need your help.
In this competition, youll develop a model to classify (and if present, segment) pneumothorax from a set of chest radiographic images. If successful, you could aid in the early recognition of pneumothoraces and save lives.
If youre up for the challenge, take a deep breath, and get started now.
Note: As specified on the Data Page, the dataset must be retrieved from Cloud Healthcare. Review this tutorial (or in pdf format) for instructions on how to do so.
Acknowledgments
SIIM Machine Learning Committee Co-Chairs, Steven G. Langer, PhD, CIIP and George Shih, MD, MS for tirelessly leading this effort and making the challenge possible in such a short period of time.
SIIM Machine Learning Committee Members for their dedication in annotating the dataset, helping to define the most useful metrics and running tests to prepare the challenge for launch.
SIIM Hackathon Committee, especially Mohannad Hussain, for their crucial technical support with data conversion.
American College of Radiology (ACR), @RadiologyACR: For Co-hosting the challenge and Co-sponsoring the Prizes
Society of Thoracic Radiology (STR), @thoracicrad: For their unparalleled expertise in adjudicating the dataset
MD.ai: For providing the annotation tool and helping with the first layer of annotations",image data,featured,Identify Pneumothorax disease in chest x-rays ,SIIMDice,siim-acr-pneumothorax-segmentation,segmentation
Prostate cANcer graDe Assessment (PANDA) Challenge,"With more than 1 million new diagnoses reported every year, prostate cancer (PCa) is the second most common cancer among males worldwide that results in more than 350,000 deaths annually. The key to decreasing mortality is developing more precise diagnostics. Diagnosis of PCa is based on the grading of prostate tissue biopsies. These tissue samples are examined by a pathologist and scored according to the Gleason grading system. In this challenge, you will develop models for detecting PCa on images of prostate tissue samples, and estimate severity of the disease using the most extensive multi-center dataset on Gleason grading yet available.
The grading process consists of finding and classifying cancer tissue into so-called Gleason patterns (3, 4, or 5) based on the architectural growth patterns of the tumor (Fig. 1). After the biopsy is assigned a Gleason score, it is converted into an ISUP grade on a 1-5 scale. The Gleason grading system is the most important prognostic marker for PCa, and the ISUP grade has a crucial role when deciding how a patient should be treated. There is both a risk of missing cancers and a large risk of overgrading resulting in unnecessary treatment. However, the system suffers from significant inter-observer variability between pathologists, limiting its usefulness for individual patients. This variability in ratings could lead to unnecessary treatment, or worse, missing a severe diagnosis.
Automated deep learning systems have shown some promise in accurately grading PCa. Recent research, including two studies independently conducted by the groups hosting this challenge, have shown that these systems can achieve pathologist-level performance. However, these systems/results were not tested with multi-center datasets at scale.
Your work here will improve on these efforts using the most extensive multi-center dataset on Gleason grading yet. The training set consists of around 11,000 whole-slide images of digitized H&E-stained biopsies originating from two centers. This is the largest public whole-slide image dataset available, roughly 8 times the size of the CAMELYON17 challenge, one of the largest digital pathology datasets and best known challenges in the field. Furthermore, in contrast to previous challenges, we are making full diagnostic biopsy images available. Using a sizable multi-center test set, graded by expert uro-pathologists, we will evaluate challenge submissions on their applicability to improve this critical diagnostic function.
Figure 1: An illustration of the Gleason grading process for an example biopsy containing prostate cancer. The most common (blue outline, Gleason pattern 3) and second most common (red outline, Gleason pattern 4) cancer growth patterns present in the biopsy dictate the Gleason score (3+4 for this biopsy), which in turn is converted into an ISUP grade (2 for this biopsy) following guidelines of the International Society of Urological Pathology. Biopsies not containing cancer are represented by an ISUP grade of 0 in this challenge.
Radboud University Medical Center and Karolinska Institute have teamed up to organize this competition in collaboration with colleagues from Tampere University. The Computational Pathology Group (CPG) of the Radboud University Medical Center is a research group that develops computer algorithms to aid clinicians. Karolinska Institutes Department of Medical Epidemiology and Biostatistics (MEB) includes an interdisciplinary research group to improve the diagnostics and treatment of prostate cancer. Together, they hope to further their existing research to make a significant impact on the healthcare of prostate cancer patients.
Challenge organizer team: Wouter Bulten, Geert Litjens, Hans Pinckaers, Peter Strm, Martin Eklund, Lars Egevad, Henrik Grnberg, Kimmo Kartasalo, Pekka Ruusuvuori, Tomi Hkkinen, Sohier Dane, Maggie Demkin.
Sponsors
The PANDA workshop at MICCAI 2020 is sponsored by ContextVision, Ibex and Google.
Using the data outside of the competition
Interested in using the PANDA dataset outside of the competition? Please read this forum post for the latest information on the embargo and the challenge paper.",image data,featured,Prostate cancer diagnosis using the Gleason grading system,QuadraticWeightedKappa,prostate-cancer-grade-assessment-(panda)-challenge,correlation
Severstal: Steel Defect Detection,"Steel is one of the most important building materials of modern times. Steel buildings are resistant to natural and man-made wear which has made the material ubiquitous around the world. To help make production of steel more efficient, this competition will help identify defects.
Severstal is leading the charge in efficient steel mining and production. They believe the future of metallurgy requires development across the economic, ecological, and social aspects of the industryand they take corporate responsibility seriously. The company recently created the countrys largest industrial data lake, with petabytes of data that were previously discarded. Severstal is now looking to machine learning to improve automation, increase efficiency, and maintain high quality in their production.
The production process of flat sheet steel is especially delicate. From heating and rolling, to drying and cutting, several machines touch flat steel by the time its ready to ship. Today, Severstal uses images from high frequency cameras to power a defect detection algorithm.
In this competition, youll help engineers improve the algorithm by localizing and classifying surface defects on a steel sheet.
If successful, youll help keep manufacturing standards for steel high and enable Severstal to continue their innovation, leading to a stronger, more efficient world all around us.",image data,featured,Can you detect and classify defects in steel?,Dice,severstal:-steel-defect-detection,segmentation
SIIM-ISIC Melanoma Classification,"Skin cancer is the most prevalent type of cancer. Melanoma, specifically, is responsible for 75% of skin cancer deaths, despite being the least common skin cancer. The American Cancer Society estimates over 100,000 new melanoma cases will be diagnosed in 2020. Its also expected that almost 7,000 people will die from the disease. As with other cancers, early and accurate detectionpotentially aided by data sciencecan make treatment more effective.
Currently, dermatologists evaluate every one of a patients moles to identify outlier lesions or ugly ducklings that are most likely to be melanoma. Existing AI approaches have not adequately considered this clinical frame of reference. Dermatologists could enhance their diagnostic accuracy if detection algorithms take into account contextual images within the same patient to determine which images represent a melanoma. If successful, classifiers would be more accurate and could better support dermatological clinic work.
As the leading healthcare organization for informatics in medical imaging, the Society for Imaging Informatics in Medicine (SIIM)s mission is to advance medical imaging informatics through education, research, and innovation in a multi-disciplinary community. SIIM is joined by the International Skin Imaging Collaboration (ISIC), an international effort to improve melanoma diagnosis. The ISIC Archive contains the largest publicly available collection of quality-controlled dermoscopic images of skin lesions.
In this competition, youll identify melanoma in images of skin lesions. In particular, youll use images within the same patient and determine which are likely to represent a melanoma. Using patient-level contextual information may help the development of image analysis tools, which could better support clinical dermatologists.
Melanoma is a deadly disease, but if caught early, most melanomas can be cured with minor surgery. Image analysis tools that automate the diagnosis of melanoma will improve dermatologists diagnostic accuracy. Better detection of melanoma has the opportunity to positively impact millions of people.",image data,featured,Identify melanoma in lesion images,MCAUC,siim-isic-melanoma-classification,auc_multiclass
IEEE's Signal Processing Society - Camera Model Identification,"Finding footage of a crime caught on tape is an investigators dream. But even with crystal clear, damning evidence, one critical question always remainsis the footage real?
Today, one way to help authenticate footage is to identify the camera that the image was taken with. Forgeries often require splicing together content from two different cameras. But, unfortunately, the most common way to do this now is using image metadata, which can be easily falsified itself.
This problem is actively studied by several researchers around the world. Many machine learning solutions have been proposed in the past: least-squares estimates of a cameras color demosaicing filters as classification features, co-occurrences of pixel value prediction errors as features that are passed to sophisticated ensemble classifiers, and using CNNs to learn camera model identification features. However, this is a problem yet to be sufficiently solved.
For this competition, the IEEE Signal Processing Society is challenging you to build an algorithm that identifies which camera model captured an image by using traces intrinsically left in the image. Helping to solve this problem would have a big impact on the verification of evidence used in criminal and civil trials and even news reporting.",image data,featured,Identify from which camera an image was taken,WeightedCategorizationAccuracy,ieees-signal-processing-society-camera-model-identification,categorization_accuracy
State Farm Distracted Driver Detection,"Weve all been there: a light turns green and the car in front of you doesnt budge. Or, a previously unremarkable vehicle suddenly slows and starts swerving from side-to-side.
When you pass the offending driver, what do you expect to see? You certainly arent surprised when you spot a driver who is texting, seemingly enraptured by social media, or in a lively hand-held conversation on their phone.
According to the CDC motor vehicle safety division, one in five car accidents is caused by a distracted driver. Sadly, this translates to 425,000 people injured and 3,000 people killed by distracted driving every year.
State Farm hopes to improve these alarming statistics, and better insure their customers, by testing whether dashboard cameras can automatically detect drivers engaging in distracted behaviors. Given a dataset of 2D dashboard camera images, State Farm is challenging Kagglers to classify each drivers behavior. Are they driving attentively, wearing their seatbelt, or taking a selfie with their friends in the backseat?",image data,featured,Can computer vision spot distracted drivers?,MulticlassLoss,state-farm-distracted-driver-detection,multiclass_log_loss
TGS Salt Identification Challenge,"Several areas of Earth with large accumulations of oil and gas also have huge deposits of salt below the surface. But unfortunately, knowing where large salt deposits are precisely is very difficult. Professional seismic imaging still requires expert human interpretation of salt bodies. This leads to very subjective, highly variable renderings. More alarmingly, it leads to potentially dangerous situations for oil and gas company drillers. To create the most accurate seismic images and 3D renderings, TGS (the world’s leading geoscience data company) is hoping Kaggle’s machine learning community will be able to build an algorithm that automatically and accurately identifies if a subsurface target is salt or not.",image data,featured,Segment salt deposits beneath the Earth's surface,IntersectionOverUnionObjectSegmentation,tgs-salt-identification-challenge,segmentation
Global Wheat Detection ,"Open up your pantry and you’re likely to find several wheat products. Indeed, your morning toast or cereal may rely upon this common grain. Its popularity as a food and crop makes wheat widely studied. To get large and accurate data about wheat fields worldwide, plant scientists use image detection of ‘wheat heads’—spikes atop the plant containing grain. These images are used to estimate the density and size of wheat heads in different varieties. Farmers can use the data to assess health and maturity when making management decisions in their fields. However, accurate wheat head detection in outdoor field images can be visually challenging. There is often overlap of dense wheat plants, and the wind can blur the photographs. Both make it difficult to identify single heads. Additionally, appearances vary due to maturity, color, genotype, and head orientation. Finally, because wheat is grown worldwide, different varieties, planting densities, patterns, and field conditions must be considered. Models developed for wheat phenotyping need to generalize between different growing environments. Current detection methods involve one- and two-stage detectors (Yolo-V3 and Faster-RCNN), but even when trained with a large dataset, a bias to the training region remains. The Global Wheat Head Dataset is led by nine research institutes from seven countries: the University of Tokyo, Institut national de recherche pour l’agriculture, l’alimentation et l’environnement, Arvalis, ETHZ, University of Saskatchewan, University of Queensland, Nanjing Agricultural University, and Rothamsted Research. These institutions are joined by many in their pursuit of accurate wheat head detection, including the Global Institute for Food Security, DigitAg, Kubota, and Hiphen. In this competition, you’ll detect wheat heads from outdoor images of wheat plants, including wheat datasets from around the globe. Using worldwide data, you will focus on a generalized solution to estimate the number and size of wheat heads. To better gauge the performance for unseen genotypes, environments, and observational conditions, the training dataset covers multiple regions. You will use more than 3,000 images from Europe (France, UK, Switzerland) and North America (Canada). The test data includes about 1,000 images from Australia, Japan, and China. Wheat is a staple across the globe, which is why this competition must account for different growing conditions. Models developed for wheat phenotyping need to be able to generalize between environments. If successful, researchers can accurately estimate the density and size of wheat heads in different varieties. With improved detection farmers can better assess their crops, ultimately bringing cereal, toast, and other favorite dishes to your table.",image data,research,Can you help identify wheat heads using image analysis?,RSNAObjectDetectionAP,global-wheat-detection,detection
Histopathologic Cancer Detection,"In this competition, you must create an algorithm to identify metastatic cancer in small image patches taken from larger digital pathology scans. The data for this competition is a slightly modified version of the PatchCamelyon (PCam) benchmark dataset (the original PCam dataset contains duplicate images due to its probabilistic sampling, however, the version presented on Kaggle does not contain duplicates).",image data,playground,Identify metastatic tissue in histopathologic scans of lymph node sections,AUC,histopathologic-cancer-detection,auc
Pneumonia Diagnosis,"Context
Pneumonia is an infection that causes inflammation in the lungs.
It will be caused by a virus, bacteria, fungi, etc.
Radiologists check chest X-ray images to look for white spots called infiltrates in the lungs, that identify an infection.
Content
The original dataset is in Kaggle datasets.
This dataset is created by modifying ""Chest X-Ray Images (Pneumonia)"" (Licensed under CC BY 4.0)
This dataset has 5,528 X-Ray images (JPEG) and 2 categories (Pneumonia/Normal).
Training Data and Test Data
The training data is randomly chosen 3,869 (70%) images , 2,991 represents pneumonia, 878 are normal.
The test data is the remaining 1,659 (30%) images, 1,282 represents pneumonia, 377 images are normal.
Reference
Paul Mooney: Chest X-Ray Images (Pneumonia), Kaggle datasets
Daniel Kermany, Kang Zhang, Michael Goldbaum: Labeled Optical Coherence Tomography (OCT) and Chest X-Ray Images for Classification, Mendeley
Radiological Society of North America, Inc.: Pneumonia, RadiologyInfo.org for patients",image data,inclass,The 9th 1056Lab Data Analytics Competition (Extra),auc,pneumonia-diagnosis,auc
car-classification,Homework competition with cars classification,image data,inclass,Car classification competition,categorizationaccuracy,car-classification,categorization_accuracy
Predict the missing pixel value v2,A small regression problem.,image data,inclass,A simple regression tasked derived from the image of my cat.,mse,predict-the-missing-pixel-value-v2,mse
Anokha AI Adept,"Dont Overfit Images!
Prove your natural intelligence by building artificial intelligence.
The last decade was the golden age of machine learning with the increase in datasets and computation power. Deeplearning is solving a wide variety of problems with large data-sets available. However there are lot of problems which doesnt have large data-sets When trained on these datasets Neural networks tend to overfit and performs badly on test data.
Can you prove machine learning can predict accurately with less overfitting on small datasets too?.
Your task
Given 500 brain scan images and class labels, you need to identify the type of brain tumor present in 2564 images in test set. This is a multiclass image classification problem where you required to classify each image into one of the 3 class labels. See the Data section for more information about the dataset. Evaluation metrics used is Mean F1Score. Check the Evaluation section for more details on it.
Click here for GuideLines on how to make a submission through kernel
During the contest, private leaderboard will show scores only on 50% of test data. Public Leaderboard will display scores on complete test data will only be available once the contest ends.
One proves his expertise in machine learning only when he masters the art of not overfitting!.
All the best!",image data,inclass,"This machine learning contest is part of Anokha, a national level TechFest",meanfscore,anokha-ai-adept,f_score_multiclass
COMP 750/850 Project 1,"This goal of this competition is to build the best model you can for classifying 28x28 grayscale glyphs of the letters A - J. This dataset is similar to the MNIST dataset, but the classification task will be more challenging.
Each submission you make will be scored based on its classification accuracy on a held-out test dataset. Results from scoring your model on 50% of the test dataset will be used during the competition to rank your submission (and accordingly, your model) on the public leaderboard. After the end of the competition, your model will be scored again on the remaining 50% of the test dataset, and these scores will be used to determine the final rankings. In this way, we verify that your model can generalize to unseen data.
You can do your work directly on Kaggle using the Notebooks feature. This will give you access to GPUs for faster computation as well.
This is a friendly competition for educational purpose and bragging rights; as an added incentive, the team with the highest-ranked model after the close of the competition will earn a 10% bonus on their project grade. See myCourses for details of project grading.",image data,inclass,Develop machine learning models to classify notMNIST images.,categorizationaccuracy,comp-750/850-project-1,categorization_accuracy
AILAB ML Training #1,"AILAB ML Training #1
Training objectives:
Understand public and private test system.
Try validation, especially CV.
Try data augmentation.
Try ensemble.
Apply the past competition knowledge into this ongoing competition.",image data,inclass,kuzushiji MNIST (KMNIST) Classification,categorizationaccuracy,ailab-ml-training-#1,categorization_accuracy
AILAB ML Training #0,"


YUMNIST () ",image data,inclass,MNIST Classification,categorizationaccuracy,ailab-ml-training-#0,categorization_accuracy
ASN10e Final Submission - Detect COML Faces,"Steps to Complete This Assignment
Each member in your team will create a new Kaggle account using his/her terpmail.umd.edu email address (if you dont already have one).
One member from your team will go to the Team tab on this page, Save Team Name as your team name on ELMS, e.g. T0-ai girls, other members will have to join the competition and Save Team Name as his/her terpmail.umd.edu email address.
In the same Team tab, under the section Invite Others Merge with other teams or invite users to your team by their team name, add the Kaggle accounts of other members using his/her terpmail.umd.edu address to your team, other members will have to accept the merge request.
Select the Notebooks tab, create a new notebook and set the notebook name as your team name on ELMS, e.g. T0-ai girls.
In the notebook, select the Share the button on the top right corner, set the privacy setting to Public, and then add each team member as Collaborators.
Use the notebook to run the code for prediction, generate the prediction file, and visualize the predictions.
In the notebook, select Save Version on the top right corner, and select Save & Run All (Commit), this will be the version of your final submission.
In the notebook, select the version number next to the Save Version button, select Go to Viewer, scroll down to the Output Files section, select Submit to Competition.
If done correctly, your submission should appear on the Leaderboard.
How Your Grade Will Be Determined
Many of the prompts have no wrong answer. You will be given full credit based on effort.
Submit Kaggle notebook with visualization: 4 points
Submit prediction file from Kaggle notebook: 1 point
Submissions
You must make submissions directly from your teams public Kaggle notebook. If done correctly, you should see your notebook and your submission score publicly by going to the Notebooks tab.
Visualization Format
Using your predicted id and bounding box labels, display the bounding boxes with persons name on the images provided.
Visualization example:
Submission Format
For each image in the Data, you must predict the vectors of (id, xmin, xmax, ymin, ymax). Each field is described as:
id:
string type, concatenate the image_id and person_id, e.g. 1.jpg & Raymond Tu -> 1_25
image_id values are based on the filename without the file extension
person_id values are based on the person_id_mapping.csv provided
xmin:
float type, left pixel value of the bounding box
xmax:
float type, right pixel value of the bounding box
ymin:
float type, top pixel value of the bounding box
ymax
float type, bottom pixel value of the bounding box
All bounding box values are based on pixel coordinates of the provided images. All predictions must be submitted as a file in CSV format. An example of the submission file is provided in sample_submission.csv.
Evaluation
In this competition, submissions are evaluated by computing the Root-mean-squared Error (RMSE) of all matches.
A match is defined as:
the prediction bounding box has an id label that exists in the ground truth.
The final RMSE is computed as the normalized distance between the vectors of predicted bounding box values and the vectors of ground truth bounding box values, i.e. RMSE(xmin, xmax, ymin, ymax) for all matches. A lower RMSE value means better prediction. Your team will be ranked on this final metric.
Additional Information
Late submissions will not be accepted for this assignment.
You are responsible for creating a valid submission.
Zero credit will be given for this assignment if your notebook or submission file cannot be opened on Kaggle.",image data,inclass,Use Your Model to Find and Recognize COML Faces in the Images Provided,rmse,asn10e-final-submission-detect-coml-faces,mse
Digit recognition,"Welcome to the Data Science Guilds first Datathon
Competition Description
MNIST (""Modified National Institute of Standards and Technology"") is the hello world dataset of computer vision. It was released in 1999, and since then it has served as the basis for benchmarking classification algorithms. As new machine learning techniques emerge, MNIST remains a reliable resource for researchers and learners alike.
In this competition, your goal is to correctly identify digits from a dataset of tens of thousands of handwritten images. This dataset is not the real MNIST data but is quite similar. We encourage you to experiment with different algorithms to learn first-hand what works well and how techniques compare.",image data,inclass,The Data Science Guild presents the first datathon - A twist to MNIST ,categorizationaccuracy,digit-recognition,categorization_accuracy
Fashion MNIST challenge2019,"

Fashion-MNIST is Zalandos42,00028,00028x28

DataOverview

 https://www.kaggle.com/kenichinakatani/fashon-mnist-with-cnn/

Acknowledgements
We thank Zalando for providing this dataset (Fashion MNIST).",image data,inclass,for IWASAKI ML CLASS ONLY,categorizationaccuracy,fashion-mnist-challenge2019,categorization_accuracy
Digit Classification DL Workshop,Just classify the digits shown in the npz file and save the results as a CSV file. Accuracy is all that matters!,image data,inclass,Classifying funky digits,categorizationaccuracy,digit-classification-dl-workshop,categorization_accuracy
GL Hack: Landmarks,Predict Facial Landmarks and win!,image data,inclass,Predict face landmarks location,mae,gl-hack:-landmarks,mae
QSTP - Deep Learning 2019,"This is the final assignment of the Deep Learning QSTP - 2019.
The task involves image classification on various types of indoor-scenery images. The total types (labels or classes) of images is 67.
You have 10934 images to Train your model.
You need to submit the predictions (the class type, an integer in the range [0,66]) for 4686 test images given in the sample_sub.csv file
The public leaderboard will evaluate the accuracy of your predictions on 2343 i.e 50% of test images.
The final rankings will be given by private leaderboard which will be based on accuracy on the remaining 50% of data. The private leaderboard will be displayed after the competition is over
Use of transfer-learning is encouraged and you can use any pre-trained model available on Pytorch here.
IMPORTANT: All the announcements from mentors, your doubts and discussions should be put up in the discussions section.
You can use the TEMPLATE NOTEBOOK given in Kernels section to start with (You need to fork this notebook) or start with your own fresh code",image data,inclass,The final assignement for QSTP Deep Learning 2019,categorizationaccuracy,qstp-deep-learning-2019,categorization_accuracy
DSNet: fastai Hackathon,"Welcome to the first DSNet Kaggle Competition + Kaggle Workshop
Before we share the competition details, a few ground rules from the community:
This competition will be a beginner friendly competition aimed at helping you get started with kaggle. Special recognition and awards will be granted to those who create useful and healthy discussions and kernels.
If youre joining from a community outside of DSNet, Click here to join DSNet slack.
The only pre-requisites for this competition are fast.ai Lecture 1-4. If you want to do a quick recap, here is L1+L2 review from our study group and L3+L4 review
No Teams are allowed for this competition, the reason is to help you build the confidence for getting started individually.
No Private discussion outside of Kaggle is allowed, for all discussions, we will be using Kaggle forums. We request you to avoid using DSNet slack for the discussions either, the idea is to get familiar with the Kaggle setup.
We will be sharing (Dirty+Misleading) Starter Kernels, please feel free to use these cautiously, note that you will have to re-engineer these approaches in order to gain a winning solution.
Schedule and How to Join
Timings for the Live Workshop: 4-6 PM IST, July 6
Join URL (Zoom Call): Link to Join
Note this will also be recorded and released on our YouTube Channel
Competition Timeline: 4PM July 6 to 8AM July 8, 2019 (All timings in IST)
Organising Team:
Aakash NS
Kartik Godawat
Prajwal Prashanth
Sanyam Bhutani
Siddhant Ujjain",image data,inclass,Competition for Data Science Network Getting Started Workshop,categorizationaccuracy,dsnet:-fastai-hackathon,categorization_accuracy
Fashion MNIST challenge201907,"

Fashion-MNIST is Zalandos42,00028,00028x28

DataOverview

Kernels( https://www.kaggle.com/kenichinakatani/makesubmission)

Acknowledgements
We thank Zalando for providing this dataset (Fashion MNIST).",image data,inclass,for IWASAKI ML CLASS ONLY,categorizationaccuracy,fashion-mnist-challenge201907,categorization_accuracy
Tobigs13_7week_competition,"Bored of MNIST
 MNIST ?    MNIST  Kannada digits !
Acknowledgements
Kaggle thanks Vinay Prabhu for providing this interesting dataset for a Playground competition. Image reference: https://www.researchgate.net/figure/speech-for-Kannada-numbers_fig2_313113588",image data,inclass,Let me do it again~,categorizationaccuracy,tobigs13_7week_competition,categorization_accuracy
University of applied sciences Mannheim,"In diesem Semesterprojekt werden Sie Straenschilder lokalisieren und klassifizieren.
Dabei werden die folgenden Klassen unterschieden:
speed limit 30 (prohibitory)
speed limit 50 (prohibitory)
speed limit 60 (prohibitory)
speed limit 70 (prohibitory)
speed limit 80 (prohibitory)
restriction ends 80 (other)
speed limit 100 (prohibitory)
speed limit 120 (prohibitory)
no overtaking (prohibitory)
no overtaking (trucks) (prohibitory)
priority at next intersection (danger)
priority road (other)
give way (other)
stop (other)
no traffic both ways (prohibitory)
no entry (other)
danger (danger)
bend right (danger)
uneven road (danger)
slippery road (danger)
construction (danger)
traffic signal (danger)
school crossing (danger)
snow (danger)
go right (mandatory)
go left (mandatory)
go straight (mandatory)
go right or go straight (mandatory)
keep right (mandatory)
roundabout (mandatory)
restriction end (overtaking (trucks)) (other)
Die Koordinaten-Schwerpunkte liegen in X und Y Pixelkoordinaten vor, die Klassen jeweils noch als Integer-Zahl.
In Ihrer Submission-File mssen die X und Y Werte als Float Values zwischen 0 und 1 angegeben werden und die Klassen als categorical (one-hot-encoding).",image data,inclass,"In this competion, students will detect german street signs",ae,university-of-applied-sciences-mannheim,mae
JAMP Hackathon Drive 1,"Below is the path on .7 to access the hackathon data.
/data/hackathon_drive/
The data set consists of two classes  Human & Non-human.
Train set consists of 750 images and test set consists of 250 approximately.
Leader board metric used is Accuracy - (TP+TN)/(TP+TN+FP+FN)
resultscript.pyc can be used to find your algorithms performance. o Command  python resultscript.pyc {pathtoyoursubmissionfile}
o The submission file MUST be is the same format as the sample_submission.csv.
The Public score is calculated on 50% of the test data and final standings will be on the other 50%
20 Submissions are possible per day.",image data,inclass,Computer Vision - Binary Classification,categorizationaccuracy,jamp-hackathon-drive-1,categorization_accuracy
UC Irvine Math 10 Winter 2020,"Overview
This is the in-class final project of Math 10 at UC Irvine for the Winter 2020 quarter.



We will use the techniques learned in class to classify the hiragana characters in Japanese. You may use any set of techniques that you like mentioned in the project page on Canvas.
Data
The dataset we will use is Kuzushiji-MNIST, which is similar to the handwritten digits dataset MNIST. Even though in the evaluation, we are using the test data, but the set has been permuted randomly, so a direct copy of the testing label file will not work. Please follow the starter code on the Canvas page/Github repo to import the data to a notebook, and generate a sample answer see if you can get the system to grade your submission.
Teams
Each team will consist at most 3 members, the selection will be randomly generated weighing the midterm after the free team sign-up phase, and will be available on Canvas group function in People tab.
Submission
Please refer to Rules: you can upload .csv files generated by the starter code on the Canvas page/Github repo to this competition. The submission is then cross-validated by using about 50% of the testing dataset. I upload a benchmark solution myself using a 5-nearest neighbor algorithm, which achieves about 90% accuracy on the testing dataset. Your solution should be at least on par with this benchmark.
Acknowledgements
We would like to thank Tarin Clanuwat for allowing us to use this dataset.",image data,inclass,Can your algorithm read ancient Japanese literature?,categorizationaccuracy,uc-irvine-math-10-winter-2020,categorization_accuracy
ISSM2020 AI Challenge,"Purpose of this event
The International Symposium on Semiconductor Production Technology (International Symposium on Semiconductor Manufacturing: ISSM) will hold the ISSM AI Technology Contest 2020 this year to utilize AI for semiconductor manufacturing site data in order to develop artificial intelligence technology (AI) in the field of semiconductor manufacturing equipment.
The purpose of this contest is to broaden the base of practical research and development using real data generated at semiconductor manufacturing equipment sites. This contest is for the ""defective particle classification in SEM images"" algorithm development, which is essential for improving the yield of semiconductor manufacturing. Participants in this contest will create a learning model that automatically classifies 4,000 particle SEM images used in actual semiconductor manufacturing into the specified category.
The First ISSM AI Solution Contests To Revolutionalize Semiconductor Manufacturing

International Symposium on Semiconductor Manufacturing: ISSM)  (AI)ISSM AI 2020AI
  SEM4,000SEM
ISSM AI2020
What skills do we want?
In addition to basic image recognition and classification methods, it is necessary to solve issues that occur during actual operation at semiconductor manufacturing sites, such as imbalanced data, unclear classes, and micro-defects.


Award
Based on the correct answer rate of the classification, the best prize and the technical prize are awarded. A certificate of commendation and a prize will be awarded to the awarding team.
Award Ceremony
The team of outstanding players will be honored at ISSM 2020 (on December 15-16, 2020).

    
ISSM 202020201215~16ONLINE",image data,inclass,SEM image classification competition,categorizationaccuracy,issm2020-ai-challenge,categorization_accuracy
notMNIST Competition,"The MNIST dataset is a classic of machine learning, in which the goal is to correctly classify 28 x 28 grayscale images of handwritten digits 0 - 9. In this competition, youre going to be working on a similar but more challenging task: correctly classifying 28 x 28 grayscale glyphs of letters A - J, a subset of the so-called notMNIST dataset. This is a great way to experiment with what youe learned about convolutional neural networks.
Acknowledgements
Thanks to Yaroslav Bulatov for creating this dataset.",image data,inclass,Correctly classify images of letters A - J in different fonts ,categorizationaccuracy,notmnist-competition,categorization_accuracy
ML Hackathon 2019 Q2,"Welcome to the ML Hackathon 2019, conducted by the Developer Students Club at ASE, Coimbatore !
This dataset contains of three fruits - apples, oranges and pomegranates.
Your task is to classify the test data images into one of these three cateogories.


While classifying the test images, please follow our notation:
0 - pomegranates
1 - oranges
2 - apples

Acknowledgements:


[1] Horea Muresan, Mihai Oltean, Fruit recognition from images using deep learning, Acta Univ. Sapientiae, Informatica Vol. 10, Issue 1, pp. 26-42, 2018.

[2] Sriram Reddy Kalluri: Fruits fresh and rotten for classification: Apples Oranges Bananas dataset

[3] Google images",image data,inclass,Train a neural network to classify 3 kinds of fruit!,categorizationaccuracy,ml-hackathon-2019-q2,categorization_accuracy
PadhAI: Hindi Vowel - Consonant Classification,"About Us
One Fourth Labs is an IIT Madras incubated startup with a goal to make India ready for the AI age. We want to skill Indian workforce in the areas of Artificial Intelligence (AI) at almost one-tenth the industry upskilling price.
Our flagship online school PadhAI provides India-specific courses on AI, and is open to all students, faculty, and professionals with a basic background in mathematics and python
Task
The goal is to identify the vowel and the consonant of each character image using Convolutional Neural Networks.
Evaluation Metric
Submissions are evaluated on Accuracy Score between the predicted and the actual labels on test dataset",image data,inclass,Can you predict the vowel and consonant of a Hindi character image?,categorizationaccuracy,padhai:-hindi-vowel-consonant-classification,categorization_accuracy
PadhAI: Tamil Vowel - Consonant Classification,"About Us
One Fourth Labs is an IIT Madras incubated startup with a goal to make India ready for the AI age. We want to skill Indian workforce in the areas of Artificial Intelligence (AI) at almost one-tenth the industry upskilling price.
Our flagship online school PadhAI provides India-specific courses on AI, and is open to all students, faculty, and professionals with a basic background in mathematics and python
Task
The goal is to identify the vowel and the consonant of each character image using Convolutional Neural Networks.
Evaluation Metric
Submissions are evaluated on Accuracy Score between the predicted and the actual labels on test dataset",image data,inclass,Can you predict the vowel and consonant of a Tamil character image?,categorizationaccuracy,padhai:-tamil-vowel-consonant-classification,categorization_accuracy
PadhAI: Text - Non Text Classification Level 4b,"About Us
One Fourth Labs is an IIT Madras incubated startup with a goal to make India ready for the AI age. We want to skill Indian workforce in the areas of Artificial Intelligence (AI) at almost one-tenth the industry upskilling price.
Our flagship online school PadhAI provides India-specific courses on AI, and is open to all students, faculty, and professionals with a basic background in mathematics and python
Task
The goal is to identify the presence of a character in images using MP Neuron / Perceptron / Perceptron with sigmoid. The character images are compiled in Tamil, Hindi and English. We have altered the task in 4 levels with increase in data complexity.
Evaluation Metric
Submissions are evaluated on Accuracy Score between the predicted and the actual labels on test dataset
Acknowledgements
Background Data:
http://www.image-net.org/
Tamil Character Data: http://www.jfn.ac.lk/index.php/data-sets-printed-tamil-characters-printed-documents/
Hindi Character Data:
https://www.kaggle.com/ashokpant/devanagari-character-dataset",image data,inclass,Can you predict whether an image has TEXT or NOT?,categorizationaccuracy,padhai:-text-non-text-classification-level-4b,categorization_accuracy
PadhAI: Text - Non Text Classification Level 4a,"About Us
One Fourth Labs is an IIT Madras incubated startup with a goal to make India ready for the AI age. We want to skill Indian workforce in the areas of Artificial Intelligence (AI) at almost one-tenth the industry upskilling price.
Our flagship online school PadhAI provides India-specific courses on AI, and is open to all students, faculty, and professionals with a basic background in mathematics and python
Task
The goal is to identify the presence of a character in images using MP Neuron / Perceptron / Perceptron with sigmoid. The character images are compiled in Tamil, Hindi and English. We have altered the task in 4 levels with increase in data complexity.
Evaluation Metric
Submissions are evaluated on Accuracy Score between the predicted and the actual labels on test dataset
Acknowledgements
Background Data:
http://www.image-net.org/
Tamil Character Data: http://www.jfn.ac.lk/index.php/data-sets-printed-tamil-characters-printed-documents/
Hindi Character Data:
https://www.kaggle.com/ashokpant/devanagari-character-dataset",image data,inclass,Can you predict whether an image has TEXT or NOT?,categorizationaccuracy,padhai:-text-non-text-classification-level-4a,categorization_accuracy
Oxford Fast AI Week 2,"Welcome to our first graded competition.
Check out the starter Kernel to get going.",image data,inclass,Graded contest: Classifying flowers,categorizationaccuracy,oxford-fast-ai-week-2,categorization_accuracy
National Data Science Challenge 2019 - Advanced,"Everyday there are hundreds of thousands new products added to Shopee.
To make relevant products easily discoverable, one fundamental challenge is to accurately extract relevant information from large volume of products. For NDSC 2019, we present this real-world challenge of building an automatic solution to extract product related information through machine learning techniques. The image data is now available. To download the image data, see the Data page.
The theme of NDSC 2019 is Product Information Extraction in the Wild - a challenge to build up:
an automatic solution to extract product related information from large volume of images and free text data.
There will be two main competitions: beginner product category classification and advanced product information extraction. Participants may enter either (or both) of these two competitions, and can choose to tackle any (or all) of the data sources provided on the Data pages: Beginner Category and Advanced Category.
On this Kaggle page, we introduce the advanced product information extraction task with advanced category information. Participants are required to detect the brand, model, and other attributes. For those who are also interested in the junior-level task using Beginner Category, please refer to Product Category Classification Kaggle page for more details.",image data,inclass,Product Information Extraction,map@{k},national-data-science-challenge-2019-advanced,reconstruction
Cats vs Dogs vs More,"As part of the Kainos AI Camp, we are running our own (small) Kaggle competition. Unlike regular Kaggle competitions, only other Campers are taking part. The aim is to achieve the highest accuracy.
How to form a team
You can have a team of up to three people. First you will have to accept the competition rules. To form a team, click on My Team from the dashboard. There you can search for and add other Kagglers to your team. Please note: until your teammates have accepted the competition rules, they will not show up here. Be sure to designate a Team Leader. If you and someone you want to work with have both formed teams, it is possible to merge teams (as long as the total number of people is below 3).
The aim of the competition
You will be given a dataset containing images, each of which belongs to one of eight different classes. These are: airplane, car, cat, dog, flower, fruit, motorbike and person. Your task is to train a classifier that will label each image correctly.
The scoring will happen on a part of the dataset that you are not given to train on, so make sure your validation accuracy is high! If you overfit on the training data, your submission will have a low score! The final score is the accuracy (technically the cross-categorical accuracy) that you achieve on the test dataset.
How to submit
You can make multiple submissions - although you are limited to 10 submissions per day. The top two submissions (as in, the two best submissions) from each team will be scored at the end. Kaggle will automatically select the top two, although you can overrule this if you choose to.
Advice
Convolutional Neural Nets are much better at classifying images than other kinds of Neural Nets. However, you can also experiment with different kinds of classifiers we have learned about. Good solutions will display some understanding of the underlying theory, so if you get stuck you may want to read some more about how CNNs work. If youre very serious about winning, you have the entire weekend to work on your solution.
Prizes
Every member of the winning team will get a mystery prize, which is being awarded at the end of the Hackathon on Saturday. Well announce the winners on the day of the submission deadline, though.",image data,inclass,There are eight classes total - the highest accuracy wins!,categorizationaccuracy,cats-vs-dogs-vs-more,categorization_accuracy
[T] PadhAI: Text - Non Text Classification Level 1,"About Us
One Fourth Labs is an IIT Madras incubated startup with a goal to make India ready for the AI age. We want to skill Indian workforce in the areas of Artificial Intelligence (AI) at almost one-tenth the industry upskilling price.
Our flagship online school PadhAI provides India-specific courses on AI, and is open to all students, faculty, and professionals with a basic background in mathematics and python
Task
The goal is to identify the presence of a character in images using MP Neuron / Perceptron / Perceptron with sigmoid. The character images are compiled in Tamil, Hindi and English. We have altered the task in 4 levels with increase in data complexity.
Evaluation Metric
Submissions are evaluated on Accuracy Score between the predicted and the actual labels on test dataset
Steps to use the predefined models
You can use the predefined models namely MPNeuron, Perceptron and PercepetronWithSigmoid in padhai.py by following the steps below:
1) Add the input folder to sys path.
import sys
sys.path.insert(0, ""../input"")
2) Import the model from padhai.py.
from padhai import MPNeuron, Perceptron, PerceptronWithSigmoid
3) Instantiate the model.
model = Perceptron()
4) Fit the model using training data.
model.fit(X_train, Y_train)
5) Predict the labels for test data.
y_pred = model.predict(X_test)
Acknowledgements
Tamil Character Data: http://www.jfn.ac.lk/index.php/data-sets-printed-tamil-characters-printed-documents/
Hindi Character Data:
https://www.kaggle.com/ashokpant/devanagari-character-dataset",image data,inclass,Can you predict whether an image has TEXT or NOT?,categorizationaccuracy,[t]-padhai:-text-non-text-classification-level-1,categorization_accuracy
[Student] Shopee Code League - Product Detection,"Background
At Shopee, we always strive to ensure the correct listing and categorization of products. For example due to the recent pandemic situation, face masks become extremely popular for both buyers and sellers, everyday we need to categorize and update a huge number of masks items. A robust product detection system will significantly improve the listing and categorization efficiency. But in the industrial field the data is always much more complicated and there exists mis-labelled images, complex background images and low resolution images, etc. The noisy and imbalanced data and multiple categories make this problem still challenging in the modern computer vision field.
Note: This page is for participants from student group!
Task
In this competition, a multiple image classification model needs to be built. There are ~100k images within 42 different categories, including essential medical tools like masks, protective suits and thermometers, home & living products like air-conditioner and fashion products like T-shirts, rings, etc. For the data security purpose the category names will be desensitized. The evaluation metrics is top-1 accuracy.",image data,inclass,Detect Real Product On E-commercial Platform,categorizationaccuracy,[student]-shopee-code-league-product-detection,categorization_accuracy
Птица или самолет?,Определите что изображено на картинке: Птица или Самолет?,image data,inclass,"Определи по картинке, кто изображён на рисунке?",auc,bird-or-aircraft-2020,auc
Классификация изображений,This is a test competition for the task of classifying objects on image.,image data,inclass,Определить тип объекта по изображению,"
meanfscorebeta",img-minor-2018,f_score
Fieldguide Challenge: Moths & Butterflies,"We are conducting the very first Fieldguide Challenge, a FGVCx challenge focusing on fine-grained classification of Lepidoptera (moths & butterflies). There are an estimated 175,000 species of Lepidoptera (leps), most of which are rarely photographed. Species identification is further complicated by distinct adult and immature stages (caterpillars), and often a lack of concrete visual diagnostics for separation between near-identical species. The goal of the competition is to push state-of-the-art classification of real world data that contains high class imbalance and high infraclass variability.
Moths and butterflies are by far the most photographed group of animals outside of birds. Having an effective CV component to identification apps would engage tens of thousands of avid citizen scientists. It would allow us to create effective monitoring programs through promotion of night-lighting for moths, which are sensitive environmental indicators. We estimate that at least 50,000 species of moths and butterflies can be identified by images, and we have citizen science groups throughout the world ready to put their smartphone cameras to use.
The Fieldguide Challenge dataset contains over 5000 species, with a combined training and testing set of over 530,000 images that have been collected and verified by multiple authors of Fieldguide. The dataset features visually-similar species, many of which are highly polymorphic, various stages in the species’ lifecycle (larvae vs adult), as well as species that are often misidentified as a moth/butterfly (e.g. caddisflies).The winning team may have the opportunity to present their results at the FGVC workshop at CVPR in Long Beach. The host would like to work with winning teams to build better models for Fieldguide.ai.
Competition Objectives Improve classification accuracy over a dataset that contains high class imbalance and high heterogeneity (caterpillar vs. adult, wild vs pinned specimens, etc. Improve error detection accuracy - what is not-a-lep*",image data,inclass,Improve classification accuracy over a dataset that contains high class imbalance and high heterogeneity,"
meanbesterroratk",fieldguide-challenge-moths-and-butterflies,mae
Птица или самолет,"В модернизированном датасете CIFAR-10 научись правильно определять, что изображено на картинке: Самолет или птица.",image data,inclass,Определи что на картинке! Птица или самолет,"
categorizationaccuracy",bird-or-aircraft,categorization_accuracy
花の画像の分類,画像を使った分類タスクのコンペです。全体のデータサイズが大きくないため扱いやすいかと思います。,image data,inclass,デイジー・たんぽぽ・バラ・ひまわり・チューリップ の画像分類,"
categorizationaccuracy",flowers-recognition,categorization_accuracy
Halite by Two Sigma - Playground Edition,"Note: This simulation is a playground competition extending the fourth season of Halite for participation. We have modified the rules to serve as a two-player game instead of four-player game. No points or medals will be awarded for this competition.
Ahoy there! Theres halite to be had and ships to be deployed! Are you ready to navigate the skies and secure your territory?
Halite by Two Sigma (""Halite"") is a resource management game where you build and control a small armada of ships. Your algorithms determine their movements to collect halite, a luminous energy source. The most halite at the end of the match wins, but its up to you to figure out how to make effective and efficient moves. You control your fleet, build new ships, create shipyards, and mine the regenerating halite on the game board.
Created by Two Sigma in 2016, more than 15,000 people around the world have participated in a Halite challenge. Players apply advanced algorithms in a dynamic, open source game setting. The strategic depth and immersive, interactive nature of Halite games make each challenge a unique learning environment.
Halite IV builds on the core game design of Halite III with a number of key changes that shift the focus of the game towards tighter competition on a smaller board. New game features include regenerating halite, shipyard creation, no more ship movement costs, and stealing halite from other players!
So dust off your halite meters and fasten your seatbelts. The fourth season of Halite is about to begin!",RL,playground,Collect the most halite during your match in space,halite,halite-by-two-sigma-playground-edition,points
Hungry Geese,"Whether it be in an arcade, on a phone, as an app, on a computer, or maybe stumbled upon in a web search, many of us have likely developed fond memories playing a version of Snake. Its addicting to control a slithering serpent and watch it grow along the grid until you make one wrong move. Then you have to try again because surely you wont make the same mistake twice!
With Hungry Geese, Kaggle has taken this classic in the video game industry and put a multi-player, simulation spin to it. You will create an AI agent to play against others and survive the longest. You must make sure your goose doesnt starve or run into other geese; its a good thing that geese love peppers, donuts, and pizzawhich show up across the board.
Extensive research exists in building Snake models using reinforcement learning, Q-learning, neural networks, and more (maybe youll use Python?). Take your grid-based reinforcement learning knowledge to the next level with this exciting new challenge!",RL,playground,Don't. Stop. Eating.,hungry_geese,hungry-geese,points
"Rock, Paper, Scissors","Rock, Paper, Scissors (sometimes called roshambo) has been a staple to settle playground disagreements or determine who gets to ride in the front seat on a road trip. The game is simple, with a balance of power. There are three options to choose from, each winning or losing to the other two. In a series of truly random games, each player would win, lose, and draw roughly one-third of games. But people are not truly random, which provides a fun opportunity for AI.
Studies have shown that a Rock, Paper, Scissors AI can consistently beat human opponents. With previous games as input, it studies patterns to understand a players tendencies. But what happens when we expand the simple Best-of-3 game to be Best-of-1000? How well can artificial intelligence perform?
In this simulation competition, you will create an AI to play against others in many rounds of this classic game. Can you find patterns to make yours win more often than it loses? Its possible to greatly outperform a random player when the matches involve non-random agents. A strong AI can consistently beat predictable AI.
This problem is fundamental to the fields of machine learning, artificial intelligence, and data compression. There are even potential applications in human psychology and hierarchical temporal memory. Warm up your hands and get ready to Rock, Paper, Scissors in this challenge.
Image acknowledgements:
Photos from The Noun Project: Rock, Paper, Scissors",RL,playground,Shoot!,rps,"rock,-paper,-scissors",points
Santa 2020 - The Candy Cane Contest,"Its the most wonderful time of the year
With the elves eating candy
Theyll feel super dandy and be of good cheer
Its the most wonderful time of the year
Its the hap-happiest season of all
When spirits are lifted the toys will be gifted
And games to enthrall!
Its the hap-happiest season of all
The party for throwing
Has snow cones aglowing
With bragging rights out on display.
So now you must plan it,
To beat the armed bandits
who keep all the candy away.
Its the most wonderful time of the year!
Morale has been low at the North Pole this year. But Santa really believes in making spirits bright! So he has planned a friendly competition among the elves to keep the Christmas cheer alive and make as many toys as possible! And the winning team gets a snow cone party!
As one of the team leaders, you know that nothing keeps your fellow elves more productive and motivated than a steady supply of candy canes! But all seven levels of the Candy Cane Forest are closed for revegetation, so the only ones available are stuck in the break room vending machines. And even though you receive free snacks on the job, the vending machines are always broken and dont always give you what you want.
Due to social distancing, only two elves can be in the break room at once. You and another team leader will take turns trying to get candy canes out of the 100 possible vending machines in the room, but each machine is unpredictable in how likely it is to work. You do know, however, that the more often you try to use a machine, the less likely it will give you a candy cane. Plus, you only have time to try 2000 times on the vending machines until you need to get back to the workshop!
If you can collect more candy canes than the other team leaders, youll surely be able to help your team win Santas contest! Try your hand at this multi-armed candy cane challenge!
Image Credit: Photos by Joanna Kosinska and Misty Ladd on Unsplash.",RL,featured,May your workdays be merry and bright,mab,santa-2020-the-candy-cane-contest,points
Halite by Two Sigma,"Ahoy there! Theres halite to be had and ships to be deployed! Are you ready to navigate the skies and secure your territory?Halite by Two Sigma (""Halite"") is a resource management game where you build and control a small armada of ships. Your algorithms determine their movements to collect halite, a luminous energy source. The most halite at the end of the match wins, but its up to you to figure out how to make effective and efficient moves. You control your fleet, build new ships, create shipyards, and mine the regenerating halite on the game board.Created by Two Sigma in 2016, more than 15,000 people around the world have participated in a Halite challenge. Players apply advanced algorithms in a dynamic, open source game setting. The strategic depth and immersive, interactive nature of Halite games make each challenge a unique learning environment. Halite IV builds on the core game design of Halite III with a number of key changes that shift the focus of the game towards tighter competition on a smaller board. New game features include regenerating halite, shipyard creation, no more ship movement costs, and stealing halite from other players! So dust off your halite meters and fasten your seatbelts. The fourth season of Halite is about to begin!",RL,featured,Collect the most halite during your match in space,Custom,halite,points
Amazon.com - Employee Access Challenge,"When an employee at any company starts work, they first need to obtain the computer access necessary to fulfill their role. This access may allow an employee to read/manipulate resources through various applications or web portals. It is assumed that employees fulfilling the functions of a given role will access the same or similar resources. It is often the case that employees figure out the access they need as they encounter roadblocks during their daily work (e.g. not able to log into a reporting portal). A knowledgeable supervisor then takes time to manually grant the needed access in order to overcome access obstacles. As employees move throughout a company, this access discovery/recovery cycle wastes a nontrivial amount of time and money.
There is a considerable amount of data regarding an employees role within an organization and the resources to which they have access. Given the data related to current employees and their provisioned access, models can be built that automatically determine access privileges as employees enter and leave roles within a company. These auto-access models seek to minimize the human involvement required to grant or revoke employee access.
Objective
The objective of this competition is to build a model, learned using historical data, that will determine an employees access needs, such that manual access transactions (grants and revokes) are minimized as the employees attributes change over time. The model will take an employees role information and a resource code and will return whether or not access should be granted.
Partners
This competition is hosted in collaboration with the IEEE International Workshop on Machine Learning for Signal Processing (MLSP 2013)",tabular data,featured,"Predict an employee's access needs, given his/her job role",AUC,amazon.com-employee-access-challenge,auc
Conway's Reverse Game of Life,"The Game of Life is a cellular automaton created by mathematician John Conway in 1970. The game consists of a board of cells that are either on or off. One creates an initial configuration of these on/off states and observes how it evolves. There are four simple rules to determine the next state of the game board, given the current state:
Any live cell with fewer than two live neighbours dies, as if by underpopulation.
Any live cell with two or three live neighbours lives on to the next generation.
Any live cell with more than three live neighbours dies, as if by overpopulation.
Any dead cell with exactly three live neighbours becomes a live cell, as if by reproduction.
These simple rules result in many interesting behaviors and have been the focus of a large body of mathematics.  As Wikipedia tells it,
Ever since its publication, Conways Game of Life has attracted much interest, because of the surprising ways in which the patterns can evolve. Life provides an example of emergence and self-organization. It is interesting for computer scientists, physicists, biologists, biochemists, economists, mathematicians, philosophers, generative scientists and others to observe the way that complex patterns can emerge from the implementation of very simple rules. The game can also serve as a didactic analogy, used to convey the somewhat counter-intuitive notion that ""design"" and ""organization"" can spontaneously emerge in the absence of a designer. For example, philosopher and cognitive scientist Daniel Dennett has used the analogue of Conways Life ""universe"" extensively to illustrate the possible evolution of complex philosophical constructs, such as consciousness and free will, from the relatively simple set of deterministic physical laws governing our own universe.
The emergence of order from simple rules begs an interesting question--what happens if we set time backwards?
This competition is an experiment to see if machine learning (or optimization, or any method) can predict the game of life in reverse.  Is the chaotic start of Life predictable from its orderly ends?  We have created many games, evolved them, and provided only the end boards. You are asked to predict the starting board that resulted in each end board. Although some people have examined this problem, it is unknown (at least, to us...) just how difficult this will be.",tabular data,playground,Reverse the arrow of time in the Game of Life,MAE,conway-s-reverse-game-of-life,mae
Bosch Production Line Performance,"A good chocolate souffl is decadent, delicious, and delicate. But, its a challenge to prepare. When you pull a disappointingly deflated dessert out of the oven, you instinctively retrace your steps to identify at what point you went wrong. Bosch, one of the worlds leading manufacturing companies, has an imperative to ensure that the recipes for the production of its advanced mechanical components are of the highest quality and safety standards. Part of doing so is closely monitoring its parts as they progress through the manufacturing processes.
Because Bosch records data at every step along its assembly lines, they have the ability to apply advanced analytics to improve these manufacturing processes. However, the intricacies of the data and complexities of the production line pose problems for current methods.
In this competition, Bosch is challenging Kagglers to predict internal failures using thousands of measurements and tests made for each component along the assembly line. This would enable Bosch to bring quality products at lower costs to the end user.",tabular data,featured,Reduce manufacturing failures,MatthewsCorrelationCoefficient,jigsaw-toxic-severity-rating,correlation
Rossmann Store Sales,"Rossmann operates over 3,000 drug stores in 7 European countries. Currently, Rossmann store managers are tasked with predicting their daily sales for up to six weeks in advance. Store sales are influenced by many factors, including promotions, competition, school and state holidays, seasonality, and locality. With thousands of individual managers predicting sales based on their unique circumstances, the accuracy of results can be quite varied.",tabular data,featured,"Forecast sales using store, promotion, and competitor data",RootMeanSquarePercentageError,rossmann-store-sales,percentage_errors
Airbnb New User Bookings,"Instead of waking to overlooked ""Do not disturb"" signs, Airbnb travelers find themselves rising with the birds in a whimsical treehouse, having their morning coffee on the deck of a houseboat, or cooking a shared regional breakfast with their hosts.
New users on Airbnb can book a place to stay in 34,000+ cities across 190+ countries. By accurately predicting where a new user will book their first travel experience, Airbnb can share more personalized content with their community, decrease the average time to first booking, and better forecast demand.
In this recruiting competition, Airbnb challenges you to predict in which country a new user will make his or her first booking. Kagglers who impress with their answer (and an explanation of how they got there) will be considered for an interview for the opportunity to join Airbnbs Data Science and Analytics team.
Wondering if youre a good fit? Check out this article on how Airbnb scaled data science to all sides of their organization, and visit their careers page for more on Airbnbs mission to create a world that inspires human connection.",tabular data,recruitment,Where will a new guest book their first travel experience?,NDCG@{K},airbnb-recruiting-new-user-bookings,reconstruction
Don't Get Kicked!,"One of the biggest challenges of an auto dealership purchasing a used car at an auto auction is the risk of that the vehicle might have serious issues that prevent it from being sold to customers. The auto community calls these unfortunate purchases ""kicks"".
Kicked cars often result when there are tampered odometers, mechanical issues the dealer is not able to address, issues with getting the vehicle title from the seller, or some other unforeseen problem. Kick cars can be very costly to dealers after transportation cost, throw-away repair work, and market losses in reselling the vehicle.
Modelers who can figure out which cars have a higher risk of being kick can provide real value to dealerships trying to provide the best inventory selection possible to their customers.
The challenge of this competition is to predict if the car purchased at the Auction is a Kick (bad buy).",tabular data,featured,Predict if a car purchased at auction is a lemon,Gini,dont-get-kicked!,auc
Cloud Faculty Institute Workshop,"Cloud Faculty Institute Workshop (6/22/2018)
Predict Diabetes from Medical Records
Kaggle InClass Competition
Kaggle Kernels",tabular data,inclass,Predict Diabetes from Medical Records,meanfscore,cloud-faculty-institute-workshop,f_score_multiclass
Atividade_3_PMR3508,"Introduo
Prezados Alunos,
Bem vindos  terceira atividade de PMR 3508! Nas ltimas duas atividades, ns nos concentramos em atividades de classificao. Desta vez, vamos trabalhar com um problema de regresso!
Objetivo
Nesta tarefa, vocs tm que construir um modelo que prev o preo mediano de uma casa em uma regio da Califrnia. A descrio completa dos dados disponveis pode ser encontrada na pgina ""Data"".
O objetivo desta tarefa  que vocs testem toda a gama de modelos de regresso que vocs j aprenderam at agora na matria - KNN, LASSO, RIDGE, rores de deciso entre outros para ver qual deles  mais adequado a esta tarefa. A idia  que vocs experimentem livremente com os algoritmos e os conheam melhor, entendendo como funcionam seus hiper-parmetros e entendendo um pouco sobre o tradeoff de vis e varincia.
Para algumas idias de algoritmos de regresso a usar, recomendo dar uma olhada em: http://scikit-learn.org/stable/supervised_learning.html
Avaliao
Assim como nas demais atividades, SUA NOTA SER DADA SOMENTE COM BASE NO KERNEL SUBMETIDO - estudantes que no submeterem seus kernels  competio tero suas notas zeradas, independentemente de quo bem tenham ido no Leaderboard. Lembramos tambm que  necessrio que O KERNEL SEJA CLASSIFICADO COMO PBLICO a fim que o monitor consiga ter acesso a ele para corrig-lo - casos de kernels publicados como privados e s depois da correo mudados para pblicos sofrero desconto na nota.
Pontuao
As notas desta atividade sero atribudas da seguitne maneira:
1- Completion) 10 pontos sero atribudos ao aluno que fizer a atividade e que apresentar claramente o estudo de ao menos 3 diferentes tipos de regressor
2- Organizao)1 ponto extra ser conferidos aos alunos cujos notebooks estiverem bem organizados e com explicaes claras
3- Explorao) 1 ponto extra ser dado aos alunos que fizerem uma explorao profunda dos dados existentes, com visualizaes ricas e anlises qualitativas dos dados observados
4- Feature Engineering) 2 pontos extras sero dados aos alunos cujo feature engineering for criativo. Embora se tenha poucas features sobre cada um dos lugares, existem vrias features complementares que podem ser criadas para auxiliar no processo regresso, como mdia de quartos por casa, mdia de pessoas por quarto entre outros. Alm disso, so dadas as coordenadas geogrficas de cada um dos locais. Ser que vocs conseguem criar alguma varivel criativa com isso, como a distncia do local at o mar ou a distncia do local at um dos plos do vale do Silcio? Be creative! Para os curiosos, a biblioteca geopy (https://github.com/geopy/geopy) pode ser muito til nesta etapa.
Competio
Para conferir a esta competio um carter mais competitivo, assim como na primeira, o aluno que obtiver o primeiro lugar na leaderboard de previso ganhar uma barra Lindt de 100g
Acknowledgements
The data for this competiton was taken from http://www.dcc.fc.up.pt/~ltorgo/Regression/DataSets.html, with special thanks to professor Lus Torgo for making it publicly available.",tabular data,inclass,Terceira atividade do curso PMR3508: regressão.,rmsle,atividade_3_pmr3508,percentage_errors
Chh-OLA,"In the Global Entrepreneurship Summit, Chh-OLA, the taxi-hailing startup led by Khanchandani, Harsh and Dwivedi garnered a lot of interest and funding from the investors. Five years later, Chh-OLA is successfully running a number of taxis on the streets of New Delhi and garnering significant profit.
Now, Chh-OLA has opened a Data Science division recently and you have been recruited into the division. As part of your training, Shruti, the head of the Data Science division, has given you a task of estimating total fare amount of trips using various trip parameters such as distance, passenger count etc.
Can you fulfil this task given by your head?",tabular data,inclass,Help the developers to predict total fare amount of trips,rmse,chh-ola,mse
[ACM] Recommender System Practice,"Demo for ACM.
Photo by Sebastian Ervi on Unsplash.",tabular data,inclass,Recommend talent that users might like!,map@{k},[acm]-recommender-system-practice,reconstruction
IIITB ML Project: SFO Crime Classification,Please refer to the San Francisco Crime Classification Kaggle competition for details.,tabular data,inclass,Predict the category of crimes that occurred in the city by the bay,multiclassloss,iiitb-ml-project:-sfo-crime-classification,multiclass_log_loss
Computational Intelligence Project,This competition is built for evaluating your results on the second phase of the CI course project.,tabular data,inclass,Upload your results for the second phase of the project.,categorizationaccuracy,computational-intelligence-project,categorization_accuracy
Categorical Feature Encoding Challenge II,"Can you find more cat in your dat?
We loved the participation and engagement with the first Cat in the Dat competition.
Because this is such a common task and important skill to master, weve put together a dataset that contains only categorical features, and includes:
binary features
low- and high-cardinality nominal features
low- and high-cardinality ordinal features
(potentially) cyclical features
This follow-up competition offers an even more challenging dataset so that you can continue to build your skills with the common machine learning task of encoding categorical variables. This challenge adds the additional complexity of feature interactions, as well as missing data.
This Playground competition will give you the opportunity to try different encoding schemes for different algorithms to compare how they perform. We encourage you to share what you find with the community.
If youre not sure how to get started, you can check out the Categorical Variables section of Kaggles Intermediate Machine Learning course.
Have Fun!",tabular data,playground,"Binary classification, with every feature a categorical (and interactions!)",AUC,categorical-feature-encoding-challenge-ii,auc
COVID19 Global Forecasting (Week 4),"This is week 4 of Kaggles COVID-19 forecasting series, following the Week 3 competition. This is the 4th competition weve launched in this series. All of the prior discussion forums have been migrated to this competition for continuity.
Background
The White House Office of Science and Technology Policy (OSTP) pulled together a coalition research groups and companies (including Kaggle) to prepare the COVID-19 Open Research Dataset (CORD-19) to attempt to address key open scientific questions on COVID-19. Those questions are drawn from National Academies of Sciences, Engineering, and Medicines (NASEM) and the World Health Organization (WHO).
The Challenge
Kaggle is launching a companion COVID-19 forecasting challenges to help answer a subset of the NASEM/WHO questions. While the challenge involves forecasting confirmed cases and fatalities between April 15 and May 14 by region, the primary goal isnt only to produce accurate forecasts. Its also to identify factors that appear to impact the transmission rate of COVID-19.
You are encouraged to pull in, curate and share data sources that might be helpful. If you find variables that look like they impact the transmission rate, please share your finding in a notebook.
As the data becomes available, we will update the leaderboard with live results based on data made available from the Johns Hopkins University Center for Systems Science and Engineering (JHU CSSE).
We have received support and guidance from health and policy organizations in launching these challenges. Were hopeful the Kaggle community can make valuable contributions to developing a better understanding of factors that impact the transmission of COVID-19.
Companies and Organizations
There is also a call to action for companies and other organizations: If you have datasets that might be useful, please upload them to Kaggles dataset platform and reference them in this forum thread. That will make them accessible to those participating in this challenge and a resource to the wider scientific community.
Acknowledgements
JHU CSSE for making the data available to the public. The White House OSTP for pulling together the key open questions. The image comes from the Center for Disease Control.
This is a Code Competition. Refer to Code Requirements for details.",tabular data,research,Forecast daily COVID-19 spread in regions around world,MCRMSLE,covid19-global-forecasting-(week-4),percentage_errors_multiclass
COVID-19 diagnostic,"Predict COVID cases!
The main goal is to predict covid-19 variable.",tabular data,inclass,Detect case using data,auc,covid-19-diagnostic,auc
Recruit Restaurant Visitor Forecasting,"Running a thriving local restaurant isnt always as charming as first impressions appear. There are often all sorts of unexpected troubles popping up that could hurt business.
One common predicament is that restaurants need to know how many customers to expect each day to effectively purchase ingredients and schedule staff members. This forecast isnt easy to make because many unpredictable factors affect restaurant attendance, like weather and local competition. Its even harder for newer restaurants with little historical data.
Recruit Holdings has unique access to key datasets that could make automated future customer prediction possible. Specifically, Recruit Holdings owns Hot Pepper Gourmet (a restaurant review service), AirREGI (a restaurant point of sales service), and Restaurant Board (reservation log management software).
In this competition, youre challenged to use reservation and visitation data to predict the total number of visitors to a restaurant for future dates. This information will help restaurants be much more efficient and allow them to focus on creating an enjoyable dining experience for their customers.",tabular data,featured,Predict how many future visitors a restaurant will receive,RMSLE,recruit-restaurant-visitor-forecasting,percentage_errors
Riiid Answer Correctness Prediction,"Riiid AIEd Challenge 2020
Challenge Website
Thank you for all those who attended the AAAI-2021 workshop on AI Education! Prize-winning teams presented their models at the AAAI-2021 Workshop on AI Education - Imagining Post-COVID Education with AI - on February 9, 2021. You can find the model write-ups on the workshop website.
Think back to your favorite teacher. They motivated and inspired you to learn. And they knew your strengths and weaknesses. The lessons they taught were based on your ability. For example, teachers would make sure you understood algebra before advancing to calculus. Yet, many students dont have access to personalized learning. In a world full of information, data scientists like you can help. Machine learning can offer a path to success for young people around the world, and you are invited to be part of this mission.
In 2018, 260 million children werent attending school. At the same time, more than half of these young students didnt meet minimum reading and math standards. Education was already in a tough place when COVID-19 forced most countries to temporarily close schools. This further delayed learning opportunities and intellectual development. The equity gaps in every country could grow wider. We need to re-think the current education system in terms of attendance, engagement, and individualized attention.
Riiid Labs, an AI solutions provider delivering creative disruption to the education market, empowers global education players to rethink traditional ways of learning leveraging AI. With a strong belief in equal opportunity in education, Riiid launched an AI tutor based on deep-learning algorithms in 2017 that attracted more than one million South Korean students. This year, the company released EdNet, the worlds largest open database for AI education containing more than 100 million student interactions.
In this competition, your challenge is to create algorithms for ""Knowledge Tracing,"" the modeling of student knowledge over time. The goal is to accurately predict how students will perform on future interactions. You will pair your machine learning skills using Riiids EdNet data.
Your innovative algorithms will help tackle global challenges in education. If successful, its possible that any student with an Internet connection can enjoy the benefits of a personalized learning experience, regardless of where they live. With your participation, we can build a better and more equitable model for education in a post-COVID-19 world.
Acknowledgements
Academic Advisors
Paul Kim, Stanford Graduate School of Education
Neil Heffernan, WPI & ASSISTments
Partners",tabular data,featured,Track knowledge states of 1M+ students in the wild,AUC,riiid-answer-correctness-prediction,auc
Give Me Some Credit,"Banks play a crucial role in market economies. They decide who can get finance and on what terms and can make or break investment decisions. For markets and society to function, individuals and companies need access to credit. 
Credit scoring algorithms, which make a guess at the probability of default, are the method banks use to determine whether or not a loan should be granted. This competition requires participants to improve on the state of the art in credit scoring, by predicting the probability that somebody will experience financial distress in the next two years.
The goal of this competition is to build a model that borrowers can use to help make the best financial decisions.
Historical data are provided on 250,000 borrowers and the prize pool is $5,000 ($3,000 for first, $1,500 for second and $500 for third).",tabular data,featured,Improve on the state of the art in credit scoring by predicting the probability that somebody will experience financial distress in the next two years.  ,AUC,give-me-some-credit,auc
Abstraction and Reasoning Challenge,"Can a computer learn complex, abstract tasks from just a few examples?
Current machine learning techniques are data-hungry and brittlethey can only make sense of patterns theyve seen before. Using current methods, an algorithm can gain new skills by exposure to large amounts of data, but cognitive abilities that could broadly generalize to many tasks remain elusive. This makes it very challenging to create systems that can handle the variability and unpredictability of the real world, such as domestic robots or self-driving cars.
However, alternative approaches, like inductive programming, offer the potential for more human-like abstraction and reasoning. The Abstraction and Reasoning Corpus (ARC) provides a benchmark to measure AI skill-acquisition on unknown tasks, with the constraint that only a handful of demonstrations are shown to learn a complex task. It provides a glimpse of a future where AI could quickly learn to solve new problems on its own. The Kaggle Abstraction and Reasoning Challenge invites you to try your hand at bringing this future into the present!
This competition is hosted by Franois Chollet, creator of the Keras neural networks library. Chollets paper on measuring intelligence provides the context and motivation behind the ARC benchmark.
In this competition, youll create an AI that can solve reasoning tasks it has never seen before. Each ARC task contains 3-5 pairs of train inputs and outputs, and a test input for which you need to predict the corresponding output with the pattern learned from the train examples.
If successful, youll help bring computers closer to human cognition and youll open the door to completely new AI applications!",tabular data,research,Create an AI capable of solving reasoning tasks it has never seen before,MeanBestErrorAtK,abstraction-and-reasoning-challenge,reconstruction
Allstate Claims Severity,"When youve been devastated by a serious car accident, your focus is on the things that matter the most: family, friends, and other loved ones. Pushing paper with your insurance agent is the last place you want your time or mental energy spent. This is why Allstate, a personal insurer in the United States, is continually seeking fresh ideas to improve their claims service for the over 16 million households they protect.
Allstate is currently developing automated methods of predicting the cost, and hence severity, of claims. In this recruitment challenge, Kagglers are invited to show off their creativity and flex their technical chops by creating an algorithm which accurately predicts claims severity. Aspiring competitors will demonstrate insight into better ways to predict claims severity for the chance to be part of Allstates efforts to ensure a worry-free customer experience.
New to Kaggle? This competition is a recruiting competition, your chance to get a foot in the door with the hiring team at Allstate.",tabular data,recruitment,How severe is an insurance claim?,MAE,allstate-claims-severity,mae
ASHRAE - Great Energy Predictor III,"Q: How much does it cost to cool a skyscraper in the summer?
A: A lot! And not just in dollars, but in environmental impact.
Thankfully, significant investments are being made to improve building efficiencies to reduce costs and emissions. The question is, are the improvements working? Thats where you come in. Under pay-for-performance financing, the building owner makes payments based on the difference between their real energy consumption and what they would have used without any retrofits. The latter values have to come from a model. Current methods of estimation are fragmented and do not scale well. Some assume a specific meter type or dont work with different building types.
In this competition, youll develop accurate models of metered building energy usage in the following areas: chilled water, electric, hot water, and steam meters. The data comes from over 1,000 buildings over a three-year timeframe. With better estimates of these energy-saving investments, large scale investors and financial institutions will be more inclined to invest in this area to enable progress in building efficiencies.
About the Host
Founded in 1894, ASHRAE serves to advance the arts and sciences of heating, ventilation, air conditioning refrigeration and their allied fields. ASHRAE members represent building system design and industrial process professionals around the world. With over 54,000 members serving in 132 countries, ASHRAE supports research, standards writing, publishing and continuing education - shaping tomorrows built environment today.
Banner photo by Federico Beccari on Unsplash",tabular data,featured,How much energy will a building consume?,RMSLE,ashrae-great-energy-predictor-iii,percentage_errors
Click-Through Rate Prediction,"In online advertising, click-through rate (CTR) is a very important metric for evaluating ad performance. As a result, click prediction systems are essential and widely used for sponsored search and real-time bidding.
For this competition, we have provided 11 days worth of Avazu data to build and test prediction models. Can you find a strategy that beats standard classification algorithms? The winning models from this competition will be released under an open-source license.",tabular data,featured,Predict whether a mobile ad will be clicked,LogLoss,click-through-rate-prediction,log_loss
Predicting a Biological Response,"The objective of the competition is to help us build as good a model as possible so that we can, as optimally as this data allows, relate molecular information, to an actual biological response.
We have shared the data in the comma separated values (CSV) format. Each row in this data set represents a molecule. The first column contains experimental data describing an actual biological response; the molecule was seen to elicit this response (1), or not (0). The remaining columns represent molecular descriptors (d1 through d1776), these are calculated properties that can capture some of the characteristics of the molecule - for example size, shape, or elemental constitution. The descriptor matrix has been normalized.",tabular data,inclass,Predict a biological response of molecules from their chemical properties,LogLoss,predicting-a-biological-response,log_loss
BNP Paribas Cardif Claims Management,"As a global specialist in personal insurance, BNP Paribas Cardif serves 90 million clients in 36 countries across Europe, Asia and Latin America.
In a world shaped by the emergence of new uses and lifestyles, everything is going faster and faster. When facing unexpected events, customers expect their insurer to support them as soon as possible. However, claims management may require different levels of check before a claim can be approved and a payment can be made. With the new practices and behaviors generated by the digital economy, this process needs adaptation thanks to data science to meet the new needs and expectations of customers.
In this challenge, BNP Paribas Cardif is providing an anonymized database with two categories of claims:
Kagglers are challenged to predict the category of a claim based on features available early in the process, helping BNP Paribas Cardif accelerate its claims process and therefore provide a better service to its customers.",tabular data,featured,Can you accelerate BNP Paribas Cardif's claims management process?,LogLoss,bnp-paribas-cardif-claims-management,log_loss
Categorical Feature Encoding Challenge,"Is there a cat in your dat?
A common task in machine learning pipelines is encoding categorical variables for a given algorithm in a format that allows as much useful signal as possible to be captured.
Because this is such a common task and important skill to master, weve put together a dataset that contains only categorical features, and includes:
binary features
low- and high-cardinality nominal features
low- and high-cardinality ordinal features
(potentially) cyclical features
This Playground competition will give you the opportunity to try different encoding schemes for different algorithms to compare how they perform. We encourage you to share what you find with the community.
If youre not sure how to get started, you can check out the Categorical Variables section of Kaggles Intermediate Machine Learning course.
Have Fun!",tabular data,playground,"Binary classification, with every feature a categorical",AUC,categorical-feature-encoding-challenge,auc
Predicting Molecular Properties,"Think you can use your data science smarts to make big predictions at a molecular level?
This challenge aims to predict interactions between atoms. Imaging technologies like MRI enable us to see and understand the molecular composition of tissues. Nuclear Magnetic Resonance (NMR) is a closely related technology which uses the same principles to understand the structure and dynamics of proteins and molecules.
Researchers around the world conduct NMR experiments to further understanding of the structure and dynamics of molecules, across areas like environmental science, pharmaceutical science, and materials science.
This competition is hosted by members of the CHemistry and Mathematics in Phase Space (CHAMPS) at the University of Bristol, Cardiff University, Imperial College and the University of Leeds. Winning teams will have an opportunity to partner with this multi-university research program on an academic publication
Your Challenge
In this competition, you will develop an algorithm that can predict the magnetic interaction between two atoms in a molecule (i.e., the scalar coupling constant).
Once the competition finishes, CHAMPS would like to invite the top teams to present their work, discuss the details of their models, and work with them to write a joint research publication which discusses an open-source implementation of the solution.
About Scalar Coupling
Using NMR to gain insight into a molecules structure and dynamics depends on the ability to accurately predict so-called scalar couplings. These are effectively the magnetic interactions between a pair of atoms. The strength of this magnetic interaction depends on intervening electrons and chemical bonds that make up a molecules three-dimensional structure.
Using state-of-the-art methods from quantum mechanics, it is possible to accurately calculate scalar coupling constants given only a 3D molecular structure as input. However, these quantum mechanics calculations are extremely expensive (days or weeks per molecule), and therefore have limited applicability in day-to-day workflows.
A fast and reliable method to predict these interactions will allow medicinal chemists to gain structural insights faster and cheaper, enabling scientists to understand how the 3D chemical structure of a molecule affects its properties and behavior.
Ultimately, such tools will enable researchers to make progress in a range of important problems, like designing molecules to carry out specific cellular tasks, or designing better drug molecules to fight disease.
Join the CHAMPS Scalar Coupling challenge to apply predictive analytics to chemistry and chemical biology.",tabular data,featured,Can you measure the magnetic interactions between a pair of atoms?,GroupMeanLogMAE,predicting-molecular-properties,mae
Conway's Reverse Game of Life 2020,"This is a relaunch of a previous competition, Conways Reverse Game of Life, with the following changes:
The grid size is larger (25 vs. 25) and the grid wraps around from top to bottom and left to right
Submissions are solved forward by the appropriate number of steps, so that any correct starting solution will achieve a maximum score. This article contains the stepping function that is used for this competition.
Obligatory Disclaimer: A lot has changed since the original competition was launched 6 years ago. With the change from ""exact starting point"" to ""any correct starting point"", it is possible to get a perfect score. We just dont know how difficult that will be. Use it as a fun learning experience, and dont spoil it for others by posting perfect solutions!
~~~~~~~~~
The Game of Life is a cellular automaton created by mathematician John Conway in 1970. The game consists of a board of cells that are either on or off. One creates an initial configuration of these on/off states and observes how it evolves. There are four simple rules to determine the next state of the game board, given the current state:
Overpopulation: if a living cell is surrounded by more than three living cells, it dies.
Stasis: if a living cell is surrounded by two or three living cells, it survives.
Underpopulation: if a living cell is surrounded by fewer than two living cells, it dies.
Reproduction: if a dead cell is surrounded by exactly three cells, it becomes a live cell.
These simple rules result in many interesting behaviors and have been the focus of a large body of mathematics. As Wikipedia states
Ever since its publication, Conways Game of Life has attracted much interest, because of the surprising ways in which the patterns can evolve. Life provides an example of emergence and self-organization. It is interesting for computer scientists, physicists, biologists, biochemists, economists, mathematicians, philosophers, generative scientists and others to observe the way that complex patterns can emerge from the implementation of very simple rules. The game can also serve as a didactic analogy, used to convey the somewhat counter-intuitive notion that ""design"" and ""organization"" can spontaneously emerge in the absence of a designer. For example, philosopher and cognitive scientist Daniel Dennett has used the analogue of Conways Life ""universe"" extensively to illustrate the possible evolution of complex philosophical constructs, such as consciousness and free will, from the relatively simple set of deterministic physical laws governing our own universe.
The emergence of order from simple rules begs an interesting questionwhat happens if we set time backwards?
This competition is an experiment to see if machine learning (or optimization, or any method) can predict the game of life in reverse. Is the chaotic start of Life predictable from its orderly ends? We have created many games, evolved them, and provided only the end boards. You are asked to predict the starting board that resulted in each end board.
This is a Code Competition. Refer to Code Requirements for details.",tabular data,playground,Reverse the arrow of time in the Game of Life,PostProcessorKernel,conways-reverse-game-of-life-2020,mae
Costa Rican Household Poverty Level Prediction,"The Inter-American Development Bank is asking the Kaggle community for help with income qualification for some of the worlds poorest families. Are you up for the challenge?
Heres the backstory: Many social programs have a hard time making sure the right people are given enough aid. Its especially tricky when a program focuses on the poorest segment of the population. The worlds poorest typically cant provide the necessary income and expense records to prove that they qualify.
In Latin America, one popular method uses an algorithm to verify income qualification. Its called the Proxy Means Test (or PMT). With PMT, agencies use a model that considers a familys observable household attributes like the material of their walls and ceiling, or the assets found in the home to classify them and predict their level of need.
While this is an improvement, accuracy remains a problem as the regions population grows and poverty declines.
To improve on PMT, the IDB (the largest source of development financing for Latin America and the Caribbean) has turned to the Kaggle community. They believe that new methods beyond traditional econometrics, based on a dataset of Costa Rican household characteristics, might help improve PMTs performance.
Beyond Costa Rica, many countries face this same problem of inaccurately assessing social need. If Kagglers can generate an improvement, the new algorithm could be implemented in other countries around the world.
This is a Kernels-Only Competition, so you must submit your code through Kernels, rather than uploading .csv predictions. You can create private Kernels and even share/edit your work with teammates by adding them as collaborators.",tabular data,playground,Can you identify which households have the highest need for social welfare assistance?,MacroFScore,costa-rican-household-poverty-level-prediction,f_score_multiclass
COVID19 Global Forecasting (Week 1),"This week 1 forecasting task is now closed for submissions. Click here to visit the week 2 version, and make a submission there.
This is one of the two complementary forecasting tasks to predict COVID-19 spread. This task is based on various regions across the world. To start on a single state-level subcomponent, please see the companion forecasting task for California, USA.
Background
The White House Office of Science and Technology Policy (OSTP) pulled together a coalition research groups and companies (including Kaggle) to prepare the COVID-19 Open Research Dataset (CORD-19) to attempt to address key open scientific questions on COVID-19. Those questions are drawn from National Academies of Sciences, Engineering, and Medicines (NASEM) and the World Health Organization (WHO).
The Challenge
Kaggle is launching two companion COVID-19 forecasting challenges to help answer a subset of the NASEM/WHO questions. While the challenge involves forecasting confirmed cases and fatalities between March 25 and April 22 by region, the primary goal isnt to produce accurate forecasts. Its to identify factors that appear to impact the transmission rate of COVID-19.
You are encouraged to pull in, curate and share data sources that might be helpful. If you find variables that look like they impact the transmission rate, please share your finding in a notebook.
As the data becomes available, we will update the leaderboard with live results based on data made available from the Johns Hopkins University Center for Systems Science and Engineering (JHU CSSE).
We have received support and guidance from health and policy organizations in launching these challenges. Were hopeful the Kaggle community can make valuable contributions to developing a better understanding of factors that impact the transmission of COVID-19.
Companies and Organizations
There is also a call to action for companies and other organizations: If you have datasets that might be useful, please upload them to Kaggles dataset platform and reference them in this forum thread. That will make them accessible to those participating in this challenge and a resource to the wider scientific community.
Acknowledgements
JHU CSSE for making the data available to the public. The White House OSTP for pulling together the key open questions. The image comes from the Center for Disease Control.
This is a Code Competition. Refer to Code Requirements for details.",tabular data,research,Forecast daily COVID-19 spread in regions around world,MCRMSLE,covid19-global-forecasting-(week-1),percentage_errors_multiclass
Forest Cover Type Prediction,"Get started on this competition with Kaggle Scripts. No data download or local environment needed!
Random forests? Cover trees? Not so fast, computer nerds. Were talking about the real thing.
In this competition you are asked to predict the forest cover type (the predominant kind of tree cover) from strictly cartographic variables (as opposed to remotely sensed data). The actual forest cover type for a given 30 x 30 meter cell was determined from US Forest Service (USFS) Region 2 Resource Information System data. Independent variables were then derived from data obtained from the US Geological Survey and USFS. The data is in raw form (not scaled) and contains binary columns of data for qualitative independent variables such as wilderness areas and soil type.
This study area includes four wilderness areas located in the Roosevelt National Forest of northern Colorado. These areas represent forests with minimal human-caused disturbances, so that existing forest cover types are more a result of ecological processes rather than forest management practices.
Acknowledgements
Kaggle is hosting this competition for the machine learning community to use for fun and practice. This dataset was provided by Jock A. Blackard and Colorado State University. We also thank the UCI machine learning repository for hosting the dataset. If you use the problem in publication, please cite:
Bache, K. & Lichman, M. (2013). UCI Machine Learning Repository. Irvine, CA: University of California, School of Information and Computer Science",tabular data,playground,Use cartographic variables to classify forest categories,CategorizationAccuracy,forest-cover-type-prediction,categorization_accuracy
Telstra Network Disruptions,"In their first recruiting competition, Telstra is
 challenging Kagglers to predict the severity of service disruptions on their network. Using a dataset of features from their service logs, youre tasked with predicting if a disruption is a momentary glitch or a total interruption of connectivity.
Telstra is on a journey to enhance the customer experience - ensuring everyone in the company is putting customers first. In terms of its expansive network, this means continuously advancing how it predicts the scope and timing of service disruptions. Telstra wants to see how you would help it drive customer advocacy by developing a more advanced predictive model for service disruptions and to help it better serve its customers.
This challenge was crafted as a simulation of the type of problem you might tackle as a member of the team at Telstra.
Kagglers who stand out will be considered for data science roles in Telstras Big Data team in Telstras absolute discretion. Highly-ranked participants will combine technical expertise and intuition in data science problems with a keen business sense and an effortless ability to work with technical and non-technical staff to turn data into real changes that impact customers. Highly-ranked participants will be considered by Telstra for interviews for employment, based on their work in the Competition and ability to meet the selection criteria for any suitable open job vacancy in Melbourne and Sydney, Australia. Participation in this Competition is not a recruitment process and Kaggle does not provide Telstra with recruitment services.",tabular data,recruitment,Predict service faults on Australia's largest telecommunications network ,MulticlassLoss,telstra-network-disruptions,multiclass_log_loss
Two Sigma Connect: Rental Listing Inquiries,"Finding the perfect place to call your new home should be more than browsing through endless listings. RentHop makes apartment search smarter by using data to sort rental listings by quality. But while looking for the perfect apartment is difficult enough, structuring and making sense of all available real estate data programmatically is even harder. Two Sigma and RentHop, a portfolio company of Two Sigma Ventures, invite Kagglers to unleash their creative engines to uncover business value in this unique recruiting competition.

Two Sigma invites you to apply your talents in this recruiting competition featuring rental listing data from RentHop. Kagglers will predict the number of inquiries a new listing receives based on the listings creation date and other features. Doing so will help RentHop better handle fraud control, identify potential listing quality issues, and allow owners and agents to better understand renters needs and preferences.

Two Sigma has been at the forefront of applying technology and data science to financial forecasts. While their pioneering advances in big data, AI, and machine learning in the financial world have been pushing the industry forward, as with all other scientific progress, they are driven to make continual progress. This challenge is an opportunity for competitors to gain a sneak peek into Two Sigmas data science work outside of finance.
Acknowledgments
This competition is co-hosted by Two Sigma and RentHop (a portfolio company of Two Sigma Ventures, which is a division of Two Sigma Investments) to encourage creativity in using real world data to solve everyday problems.",tabular data,recruitment,How much interest will a new rental listing on RentHop receive?,MulticlassLoss,two-sigma-connect:-rental-listing-inquiries,multiclass_log_loss
Test Competition Please Ignore,,tabular data,inclass,Time flies like an arrow. Fruit flies like banana.,categorizationaccuracy,test-competition-please-ignore,categorization_accuracy
Diabetes Classification,"Pima Indians Diabetes Database
Number of Instances: 768
Number of Attributes: 8 plus class
Columns Description:
Number of times pregnant
Plasma glucose concentration a 2 hours in an oral glucose tolerance test
Diastolic blood pressure (mm Hg)
Triceps skin fold thickness (mm)
2-Hour serum insulin (mu U/ml)
Body mass index (weight in kg/(height in m)^2)
Diabetes pedigree function
Age (years)
Class variable (0 or 1)
Class Distribution: (class value 1 is interpreted as ""tested positive for
diabetes"")
0 --- 500
1 ---- 268",tabular data,inclass,Predict the given person is suffering from diabetes.,categorizationaccuracy,diabetes-classification,categorization_accuracy
Regression Evaluative Lab,"This is your first evaluative lab.
You have to predict the Average Price!",tabular data,inclass,This is your first evaluative lab and it is a regression problem!,mse,regression-evaluative-lab,mse
FIA ML T5,"As a credit company, it is important to know beforehand who is able to pay their loans and who is not. The goal of this puzzle is to build a statistical/machine learning model to figure out which clients are able to honor their debt.",tabular data,inclass,Machine Learning,auc,fia-ml-t5,auc
Predicting user conversions,"75.06/95.58 Organizacin de Datos
Segundo Cuatrimestre de 2018
Trabajo Prctico 2: Enunciado
El segundo trabajo prctico es una competencia de Machine Learning en donde cada grupo debe intentar determinar, para cada usuario presentado, cul es la probabilidad de que ese usuario realice una conversin en Trocafone en un periodo determinado.",tabular data,inclass,Estimar la probabilidad de que un usuario concrete una compra.,auc,predicting-user-conversions,auc
Data-Driven Business Analytics,"Data-Driven Business Analytics
Erster Anwendungsfall
In dieser Herausforderung sehen Sie sich damit konfrontiert, den Preis von Gebrauchtwagen zu prognostizieren.",tabular data,inclass,WS 19/20 3. Projekt,r2score,data-driven-business-analytics,r2_score
Python Class - Practice,"Practice of Data Analysis - For New Generation & Intern
This is the first challenge for Intern & New Generation.

Data analysis is an important skill for everyone. Especially, AI, Machine Learning and Deep Learing are better techniques for data analysis.

You will learn more and more while participating this competition. If you have any problem, you can write something in Discussion. Also, we are willing to see you share your code & idea in Kernels.",tabular data,inclass,Data Analysis - Practice For New Generation & Intern,auc,python-class-practice,auc
EC524: Heart-disease classification,"Goal: The goal of this project is to apply everything weve covered so far in a new context: classification of heart disease. Weve covered many of the central components of machine learning and several algorithms. Now you apply them.
Specifically, you will use a set of 14 features to predict a single outcome: whether or not an individual patient has heart disease.
This competition will help introduce you to some of the differences (and the similarities) between classification problems and regression problems.",tabular data,inclass,Classify individuals' risk for heart disease,meanfscore,ec524:-heart-disease-classification,f_score_multiclass
House Price Prediction,"This competititon is hosted by data analysis club IIT Palakkad. The challenge is to predict the price of each house given some information related to houses. Goal of competition is to make you familiar with environment of kaggle and basics of regression. Go through the data page and evaluation page for further details.
For submitting predictions click Submit Predictions. Maximum submission per day is 5.",tabular data,inclass,predict the price of house,rmse,house-price-prediction,mse
DM-Assignment 1,"This is the home page of the competition. As the description of the competition suggests you are given a dataset with 199 attributes (including ID and Class) and 13000 instances and your task is to classify the given dataset into 5 classes [1,2,3,4,5] using only Clustering algorithms.",tabular data,inclass,Classify by Clustering!,categorizationaccuracy,dm-assignment-1,categorization_accuracy
Focus start 2020,      15 ,tabular data,inclass,Предсказание задержек рейсов более чем на 15 минут,auc,focus-start-2020,auc
Heart Disease Prediction,"Introduction
Machine Learning is used across many spheres around the world. The healthcare industry is no exception. Machine Learning can play an essential role in predicting the presence/absence of Locomotor disorders, Heart diseases, and more. Such information, if predicted well in advance, can provide important insights to doctors who can then adapt their diagnosis and treatment per-patient basis.
In this competition, your task is to use labeled data to train a machine-learning algorithm to be able to predict on unseen test data.
There are 13 predictor attributes as described in the Data page.
All the best. This is the place where all the knowledge learned so far would come to application.",tabular data,inclass,Predict whether a patient had a heart disease,categorizationaccuracy,heart-disease-prediction,categorization_accuracy
GirlsGoIT competition 2020,"Prezicerea popularitii cntecelor de pe Spotify
n aceast competiie va trebui s prezicei popularitatea unui cntec de pe Spotify analiznd datele din baza de date propus. Pe baza a ceea ce ai nvat la GirlsGoIT summer camp, va trebui s creai un model de clasificare/regresie care s prezic pentru fiecare cntec popularitatea - de la 1 la 10
Reguli:
Competiia se adreseaz doar participantelor la tabra de var GirlsGoIT 2020
Codul trebuie s fie scris n **Python**, alte limbaje de programare nu se accept!!!
Se vor utiliza doar modelele nvate la tabra de var GirlsGoIT!
Nu copiai ideile altor echipe, s-ar putea ca ideea voastr s fie mai bun!
Vei ncrca doar fiierul cu predicii predictions.csv, iar prezentrile le vei trimite separat trainerilor GirlsGoIT!
Evaluarea finala
Evaluarea finala va consta n rezultatul obinut la aceast competiie pe kaggle i evaluarea prezentrii powerpoint, n care vei include descrierea proiectului, grafice, etc.",tabular data,inclass,Competition for the GirlsGoIT Data Science Summer Camp 2020,categorizationaccuracy,girlsgoit-competition-2020,categorization_accuracy
Find me that fish,On eclass,tabular data,inclass,Predict the Probability of Occurence for the Marine Species Engraulis Encrasicolus,rmse,find-me-that-fish,mse
Don't Overfit! II,"Long ago, in the distant, fragrant mists of time, there was a competition
It was not just any competition.
It was a competition that challenged mere mortals to model a 20,000x200 matrix of continuous variables using only 250 training samples without overfitting.
Data scientists  including Kaggles very own Will Cukierski  competed by the hundreds. Legends were made. (Will took 5th place, and eventually ended up working at Kaggle!) People overfit like crazy. It was a Kaggle-y, data science-y madhouse.
So were doing it again.
Dont Overfit II: The Overfittening
This is the next logical step in the evolution of weird competitions. Once again we have 20,000 rows of continuous variables, and a mere handful of training samples. Once again, we challenge you not to overfit. Do your best, model without overfitting, and add, perhaps, to your own legend.
In addition to bragging rights, the winner also gets swag. Enjoy!
Acknowledgments
We hereby salute the hard work that went into the original competition, created by Phil Brierly. Thank you!",tabular data,playground,A Fistful of Samples,AUC,dont-overfit!-ii,auc
Expedia Hotel Recommendations,"Planning your dream vacation, or even a weekend escape, can be an overwhelming affair. With hundreds, even thousands, of hotels to choose from at every destination, its difficult to know which will suit your personal preferences. Should you go with an old standby with those pillow mints you like, or risk a new hotel with a trendy pool bar? 
Expedia wants to take the proverbial rabbit hole out of hotel search by providing personalized hotel recommendations to their users. This is no small task for a site with hundreds of millions of visitors every month!
Currently, Expedia uses search parameters to adjust their hotel recommendations, but there arent enough customer specific data to personalize them for each user. In this competition, Expedia is challenging Kagglers to contextualize customer data and predict the likelihood a user will stay at 100 different hotel groups.
The data in this competition is a random selection from Expedia and is not representative of the overall statistics. ",tabular data,featured,Which hotel type will an Expedia customer book?,MAP@{K},expedia-hotel-recommendations,reconstruction
Google Analytics Customer Revenue Prediction,"The 80/20 rule has proven true for many businessesonly a small percentage of customers produce most of the revenue. As such, marketing teams are challenged to make appropriate investments in promotional strategies.
RStudio, the developer of free and open tools for R and enterprise-ready products for teams to scale and share work, has partnered with Google Cloud and Kaggle to demonstrate the business impact that thorough data analysis can have.
In this competition, youre challenged to analyze a Google Merchandise Store (also known as GStore, where Google swag is sold) customer dataset to predict revenue per customer. Hopefully, the outcome will be more actionable operational changes and a better use of marketing budgets for those companies who choose to use data analysis on top of GA data.",tabular data,featured,Predict how much GStore customers will spend,RMSE,google-analytics-customer-revenue-prediction,mse
"Ghouls, Goblins, and Ghosts... Boo!","Get out your dowsing rods, electromagnetic sensors,  and gradient boosting machines. Kaggle is haunted and we need your help. After a month of making scientific observations and taking careful measurements, weve determined that 900 ghouls, ghosts, and goblins are infesting our halls and frightening our data scientists. When trying garlic, asking politely, and using reverse psychology didnt work, it became clear that machine learning is the only answer to banishing our unwanted guests.
So now the hour has come to put the data weve collected in your hands. Weve managed to identify 371 of the ghastly creatures, but need your help to vanquish the rest. And only an accurate classification algorithm can thwart them. Use bone length measurements, severity of rot, extent of soullessness, and other characteristics to distinguish (and extinguish) the intruders. Are you ghost-busters up for the challenge?",tabular data,playground,Can you classify monsters haunting Kaggle?,categorizationaccuracy,"ghouls,-goblins,-and-ghosts...-boo!",categorization_accuracy
The Winton Stock Market Challenge,"Do you laugh (and then get down to work) in the face of terabytes of noisy, non-stationary data? Winton Capital is looking for data scientists who excel at finding the hidden signal in the proverbial haystack, and who are excited by creating novel statistical modelling and data mining techniques. 
In this recruiting competition, Winton challenges you to take on the very difficult task of predicting the future (stock returns). Given historical stock performance and a host of masked features, can you predict intra and end of day returns without being deceived by all the noise? 
Research scientists at Winton have crafted this competition to be challenging and fun for the community while providing a taste of the types of problems they work on everyday. Theyre excited to connect with Kagglers who bring a unique background and creative approach to the competition.
Winton is offering cash prizes to winning teams as a reward for their work, but the intent of the competition is not commercial. The intellectual property you create remains your own and will be evaluated in the context of suitability for employment. 
For more on the culture at Winton, check out the About Winton page or their careers page.",tabular data,featured,Join a multi-disciplinary team of research scientists,WMAE,the-winton-stock-market-challenge,mae
Titanic - Machine Learning from Disaster," Ahoy, welcome to Kaggle! Youre in the right place.
This is the legendary Titanic ML competition  the best, first challenge for you to dive into ML competitions and familiarize yourself with how the Kaggle platform works.
The competition is simple: use machine learning to create a model that predicts which passengers survived the Titanic shipwreck.
Read on or watch the video below to explore more details. Once youre ready to start competing, click on the ""Join Competition button to create an account and gain access to the competition data. Then check out Alexis Cooks Titanic Tutorial that walks you through step by step how to make your first submission!
The Challenge
The sinking of the Titanic is one of the most infamous shipwrecks in history.
On April 15, 1912, during her maiden voyage, the widely considered unsinkable RMS Titanic sank after colliding with an iceberg. Unfortunately, there werent enough lifeboats for everyone onboard, resulting in the death of 1502 out of 2224 passengers and crew.
While there was some element of luck involved in surviving, it seems some groups of people were more likely to survive than others.
In this challenge, we ask you to build a predictive model that answers the question: what sorts of people were more likely to survive? using passenger data (ie name, age, gender, socio-economic class, etc).
Recommended Tutorial
We highly recommend Alexis Cooks Titanic Tutorial that walks you through making your very first submission step by step.
Overview of How Kaggles Competitions Work
1. Join the Competition
Read about the challenge description, accept the Competition Rules and gain access to the competition dataset.
2. Get to Work
Download the data, build models on it locally or on Kaggle Kernels (our no-setup, customizable Jupyter Notebooks environment with free GPUs) and generate a prediction file.
3. Make a Submission
Upload your prediction as a submission on Kaggle and receive an accuracy score.
4. Check the Leaderboard
See how your model ranks against other Kagglers on our leaderboard.
5. Improve Your Score
Check out the discussion forum to find lots of tutorials and insights from other competitors.
Kaggle Lingo Video
You may run into unfamiliar lingo as you dig into the Kaggle discussion forums and public notebooks. Check out Dr. Rachael Tatmans video on Kaggle Lingo to get up to speed!
What Data Will I Use in This Competition?
In this competition, youll gain access to two similar datasets that include passenger information like name, age, gender, socio-economic class, etc. One dataset is titled train.csv and the other is titled test.csv.
Train.csv will contain the details of a subset of the passengers on board (891 to be exact) and importantly, will reveal whether they survived or not, also known as the ground truth.
The test.csv dataset contains similar information but does not disclose the ground truth for each passenger. Its your job to predict these outcomes.
Using the patterns you find in the train.csv data, predict whether the other 418 passengers on board (found in test.csv) survived.
Check out the Data tab to explore the datasets even further. Once you feel youve created a competitive model, submit it to Kaggle to see where your model stands on our leaderboard against other Kagglers.
How to Submit your Prediction to Kaggle
Once youre ready to make a submission and get on the leaderboard:
1. Click on the Submit Predictions button
2. Upload a CSV file in the submission file format. Youre able to submit 10 submissions a day.
Submission File Format:
You should submit a csv file with exactly 418 entries plus a header row. Your submission will show an error if you have extra columns (beyond PassengerId and Survived) or rows.
The file should have exactly 2 columns:
PassengerId (sorted in any order)
Survived (contains your binary predictions: 1 for survived, 0 for deceased)
Got it! Im ready to get started. Where do I get help if I need it?
For Competition Help: Titanic Discussion Forum
Technical Help: Kaggle Contact Us Page
Kaggle doesnt have a dedicated support team so youll typically find that you receive a response more quickly by asking your question in the appropriate forum. The forums are full of useful information on the data, metric, and different approaches. We encourage you to use the forums often. If you share your knowledge, youll find that others will share a lot in turn!
A Last Word on Kaggle Notebooks
As we mentioned before, Kaggle Notebooks is our no-setup, customizable, Jupyter Notebooks environment with free GPUs and a huge repository of community published data & code.
In every competition, youll find many Kernels publically shared with incredible insights. Its an invaluable resource worth becoming familiar with. Check out this competitions Kernels here.
Ready to Compete? Join the Competition Here!",tabular data,getting started,Start here! Predict survival on the Titanic and get familiar with ML basics,categorizationaccuracy,titanic-machine-learning-from-disaster,categorization_accuracy
Higgs Boson Machine Learning Challenge,"Discovery of the long awaited Higgs boson was announced July 4, 2012 and confirmed six months later. 2013 saw a number of prestigious awards, including a Nobel prize. But for physicists, the discovery of a new particle means the beginning of a long and difficult quest to measure its characteristics and determine if it fits the current model of nature.
A key property of any particle is how often it decays into other particles. ATLAS is a particle physics experiment taking place at the Large Hadron Collider at CERN that searches for new particles and processes using head-on collisions of protons of extraordinarily high energy. The ATLAS experiment has recently observed a signal of the Higgs boson decaying into two tau particles, but this decay is a small signal buried in background noise. 
The goal of the Higgs Boson Machine Learning Challenge is to explore the potential of advanced machine learning methods to improve the discovery significance of the experiment. No knowledge of particle physics is required. Using simulated data with features characterizing events detected by ATLAS, your task is to classify events into ""tau tau decay of a Higgs boson"" versus ""background."" 
The winning method may eventually be applied to real data and the winners may be invited to CERN to discuss their results with high energy physicists.
Acknowledgements 
This competition is brought to you by

Additional support from:",tabular data,featured,Use the ATLAS experiment to identify the Higgs boson,custom metric,higgs-boson-machine-learning-challenge,significance
Grupo Bimbo Inventory Demand,"Planning a celebration is a balancing act of preparing just enough food to go around without being stuck eating the same leftovers for the next week. The key is anticipating how many guests will come. Grupo Bimbo must weigh similar considerations as it strives to meet daily consumer demand for fresh bakery products on the shelves of over 1 million stores along its 45,000 routes across Mexico.
Currently, daily inventory calculations are performed by direct delivery sales employees who must single-handedly predict the forces of supply, demand, and hunger based on their personal experiences with each store. With some breads carrying a one week shelf life, the acceptable margin for error is small.
In this competition, Grupo Bimbo invites Kagglers to develop a model to accurately forecast inventory demand based on historical sales data. Doing so will make sure consumers of its over 100 bakery products arent staring at empty shelves, while also reducing the amount spent on refunds to store owners with surplus product unfit for sale.",tabular data,featured,Maximize sales and minimize returns of bakery goods,RMSLE,grupo-bimbo-inventory-demand,percentage_errors
Homesite Quote Conversion,"Before asking someone on a date or skydiving, its
important to know your likelihood of success. The same goes for quoting home insurance prices to a potential customer. Homesite, a leading provider of homeowners insurance, does not currently have a dynamic conversion rate model that can give them confidence a quoted price will lead to a purchase. 
Using an anonymized database of information on customer and sales activity, including property and coverage information, Homesite is challenging you to predict which customers will purchase a given quote. Accurately predicting conversion would help Homesite better understand the impact of proposed pricing changes and maintain an ideal portfolio of customer segments. ",tabular data,featured,Which customers will purchase a quoted insurance plan?,AUC,homesite-quote-conversion,auc
Walmart Recruiting - Store Sales Forecasting,"One challenge of modeling retail data is the need to make decisions based on limited history. If Christmas comes but once a year, so does the chance to see how strategic decisions impacted the bottom line.
In this recruiting competition, job-seekers are provided with historical sales data for 45 Walmart stores located in different regions. Each store contains many departments, and participants must project the sales for each department in each store. To add to the challenge, selected holiday markdown events are included in the dataset. These markdowns are known to affect sales, but it is challenging to predict which departments are affected and the extent of the impact.
Want to work in a great environment with some of the worlds largest data sets? This is a chance to display your modeling mettle to the Walmart hiring teams.
This competition counts towards rankings & achievements.  If you wish to be considered for an interview at Walmart, check the box ""Allow host to contact me"" when you make your first entry. 
You must compete as an individual in recruiting competitions. You may only use the provided data to make your predictions.",tabular data,recruitment,Use historical markdown data to predict store sales,WMAE,walmart-recruiting-store-sales-forecasting,mae
NBA Rookies,"Description
Dataset for practicing classification -use NBA rookie stats to predict if player will last 5 years in league?
Practice Skills
Machine Learning fundamentals
Binary classification
Python or R basics",tabular data,inclass,Predict Career for NBA Rookies,categorizationaccuracy,nba-rookies,categorization_accuracy
IESB Norte - IGM - Maio 2019,"Os dados so do IGM, e iremos prever se a nota de matemtica de um municpio na prova de matemtica do ENEM est acima ou abaixo da mediana Brasil.
As notas sero calculadas de acordo com os seguintes critrios:
Nota mais alta na competio
Melhor Kernel de anlise exploratria na competio
Melhor Kernel de anlise de feature importances
Cronograma:
20/05 - Aula 06 - Elaborao de Modelo do Kaggle
27/05 - Aula 07 - Apresentao dos modelos",tabular data,inclass,ENEM mathematics grade prediction,categorizationaccuracy,iesb-norte-igm-maio-2019,categorization_accuracy
IESB Sul - IGM - Maio 2019,"Os dados so do IGM, e iremos prever se a nota de matemtica de um municpio na prova de matemtica do ENEM est acima ou abaixo da mediana Brasil (0  abaixo da mediana e 1  acima).
As notas sero calculadas de acordo com os seguintes critrios:
Nota mais alta na competio
Melhor Kernel de anlise exploratria na competio
Melhor Kernel de anlise de feature importances
O melhor trabalho em cada um dos 3 critrios acima ir ser premiado. com um vale R$50 no Uber Eats para gastar durante a ltima aula do curso.
Cronograma:
21/05 - Aula 06 - Elaborao de Modelo do Kaggle
28/05 - Aula 07 - Finalizar Modelo do Kaggle (apresentao dos melhores trabalhos)",tabular data,inclass,ENEM mathematics grade prediction,categorizationaccuracy,iesb-sul-igm-maio-2019,categorization_accuracy
IA1819,"Objetivo
Aplicar los conceptos vistos en las prcticas de Aprendizaje Automtico de la asignatura de Inteligencia Artificial en la resolucin de una aplicacin prctica, en el entorno de una competicin, acogida en el entorno Kaggle, en la que cada estudiante intentar obtener la mejor puntuacin.",tabular data,inclass,Entregable 2 de la asignatura IA curso 18/19,f_{beta},ia1819,f_score
House pricing,"Briefing
Today you need to predict the house prices in the USA. You can only use linear models to solve this challenge. Use of raw features is not the best idea, but you may create your own ones. Be creative and remember what you were taught!",tabular data,inclass,House sales in USA,mape,house-pricing,percentage_errors
Team ISTE's Datathon,"Data analytics capabilities enables us to analyze data with greater depth, sophistication and efficiencies through innovations such as artificial intelligence, machine learning, natural learning processing and bots.
Indian Society for Technical Education Students Chapter NIT Hamirpur has launched DATATHON, which invites participants from NIT Hamirpur, who will join us in our quest of developing models for the appropriate sourcing and usage of data across businesses. Participants are required to implement OHLCV time series analysis on the given dataset. The prizes will be given upon producing Identity card of NIT Hamirpur.
What is OHLCV?
The columns of the datasets are:-
O: Open Value
H: High Value
L: Low Value
C: Close Value
V: Volume Sell

To read more about this type of data-sets refer here.
For reference and clear understanding, sample code is provided here.",tabular data,inclass,First-ever Data Science Competition Held at NIT Hamirpur,rmse,team-istes-datathon,mse
Web Traffic Time Series Forecasting,"This competition focuses on the problem of forecasting the future values of multiple time series, as it has always been one of the most challenging problems in the field. More specifically, we aim the competition at testing state-of-the-art methods designed by the participants, on the problem of forecasting future web traffic for approximately 145,000 Wikipedia articles.
Sequential or temporal observations emerge in many key real-world problems, ranging from biological data, financial markets, weather forecasting, to audio and video processing. The field of time series encapsulates many different problems, ranging from analysis and inference to classification and forecast. What can you do to help predict future views?
This competition will run as two stages and involves prediction of actual future events. There will be a training stage during which the leaderboard is based on historical data, followed by a stage where participants are scored on real future events.
You have complete freedom in how to produce your forecasts: e.g. use of univariate vs multi-variate models, use of metadata (article identifier), hierarchical time series modeling (for different types of traffic), data augmentation (e.g. using Google Trends data to extend the dataset), anomaly and outlier detection and cleaning, different strategies for missing value imputation, and many more types of approaches.
We thank Google Inc. and Voleon for sponsorship of this competition, and Oren Anava and Vitaly Kuznetsov for organizing it.

Kaggle is excited to partner with research groups to push forward the frontier of machine learning. Research competitions make use of Kaggles platform and experience, but are largely organized by the research groups data science team. Any questions or concerns regarding the competition data, quality, or topic will be addressed by them.",tabular data,research,Forecast future traffic to Wikipedia pages,SMAPE,web-traffic-time-series-forecasting,percentage_errors
VSU ML 1 Regression,-,tabular data,inclass,Учимся решать задачу регрессии и делать EDA,rmse,vsu-ml-1-regression,mse
hackStat 2.0,"This challenge serves as the first round competition for hackStat 2.0.
In this competition you will work with a challenging dataset consisting of data of visitors of a website.
Predict the class of the type of customer as to whether the customer would be a revenue generating customer or not, by using the revenue variable as the dependent variable. The rest of the variables would be independent variables.
Upload the predicted outcome of the test set according to the format provided, to obtain the accuracy of the prediction.
Please find the details of the data and the competition in the email you received after the registrations.By solving this competition you will be able to apply and enhance your data science skills.
Good Luck!",tabular data,inclass,hackStat 2.0 - First round competitions,categorizationaccuracy,hackstat-2.0,categorization_accuracy
UI DS Summer School,"UI Data Science Summer School 2019
Halaman ini akan menjadi acuan untuk presentasi tim dan tempat untuk submit pekerjaan peserta.",tabular data,inclass,UI Data Science Summer School,rmse,ui-ds-summer-school,mse
Great Energy Predictor Shootout I,"This is an approximate replica of the 1993 energy use prediction competition run ASHRAE. Experience historic data formats and file sizes without needing to send five inch floppy disks in the mail! Were hosting this as a for-fun companion to the current featured ASHRAE competition
Special thanks to Jeff Harberl of the Texas A&M for sharing the data and an impressive act of digital record keeping!
Banner photo by Karsten Wrth (@karsten.wuerth) on Unsplash",tabular data,inclass,Replica of the original 1993 competition,rmse,great-energy-predictor-shootout-i,mse
YKC-2nd,"
multiclass
metric : f1 micro
NLP
7/719:00
7/1419:00
7/1419:00
private:public = 50:50
trainpublic, private
submit20
2


word embedding

",tabular data,inclass,YJKC-2nd,meanfscore,ykc-2nd,f_score_multiclass
Who is a Friend?,"Predict Friends or Not.
In this machine learning challenge you are to predict whether two persons are friends or not.
You are given a data of recorded events from a group of people and using the given data you have to predict whether they are friends or not.
The personal information have been masked for obvious reasons.
Safe Coding!",tabular data,inclass,Predict whether two persons are friends or not based on their meetings.,categorizationaccuracy,who-is-a-friend?,categorization_accuracy
YKC-cup-1st,store   log(),tabular data,inclass,YKC-cup-1st,rmse,ykc-cup-1st,mse
VSB Power Line Fault Detection,"Medium voltage overhead power lines run for hundreds of miles to supply power to cities. These great distances make it expensive to manually inspect the lines for damage that doesnt immediately lead to a power outage, such as a tree branch hitting the line or a flaw in the insulator. These modes of damage lead to a phenomenon known as partial discharge  an electrical discharge which does not bridge the electrodes between an insulation system completely. Partial discharges slowly damage the power line, so left unrepaired they will eventually lead to a power outage or start a fire.
Your challenge is to detect partial discharge patterns in signals acquired from these power lines with a new meter designed at the ENET Centre at VB. Effective classifiers using this data will make it possible to continuously monitor power lines for faults.
ENET Centre researches and develops renewable energy resources with the goal of reducing or eliminating harmful environmental impacts. Their efforts focus on developing technology solutions around transportation and processing of energy raw materials.
By developing a solution to detect partial discharge youll help reduce maintenance costs, and prevent power outages.",tabular data,featured,Can you detect faults in above-ground electrical lines?,MatthewsCorrelationCoefficient,vsb-power-line-fault-detection,correlation
Walmart Recruiting: Trip Type Classification,"Walmart uses both art and science to continually make progress on their core mission of better understanding and serving their customers. One way Walmart is able to improve customers shopping experiences is by segmenting their store visits into different trip types. 
Whether theyre on a last minute run for new puppy supplies or leisurely making their way through a weekly grocery list, classifying trip types enables Walmart to create the best shopping experience for every customer.
Currently, Walmarts trip types are created from a combination of existing customer insights (""art"") and purchase history data (""science""). In their third recruiting competition, Walmart is challenging Kagglers to focus on the (data) science and classify customer trips using only a transactional dataset of the items theyve purchased. Improving the science behind trip type classification will help Walmart refine their segmentation process.
Walmart is hosting this competition to connect with data scientists who break the mold.",tabular data,recruitment,Use market basket analysis to classify shopping trips,MulticlassLoss,walmart-recruiting:-trip-type-classification,multiclass_log_loss
Web Enthusiasts' Club NITK Recruitment,"Web Enthusiasts Club Recruitment Test 2018
This contest is being conducted as a part of the recruitment process for the Intelligence Group at Web Enthusiasts Club NITK.
Participants who can improve on the baseline will be shortlisted for interviews.
Getting Started
Here are a few resources:
Numpy Documentation
Pandas Documentation
Scikit-Learn Documentation
Tutorial on Random Forest Classifier",tabular data,inclass,Official contest for Web Club NITK's Intelligence Group Recruitments for 2019-20 academic year,auc,web-enthusiasts-club-nitk-recruitment,auc
Home Credit Default Risk,"Many people struggle to get loans due to insufficient or non-existent credit histories. And, unfortunately, this population is often taken advantage of by untrustworthy lenders.
Home Credit strives to broaden financial inclusion for the unbanked population by providing a positive and safe borrowing experience. In order to make sure this underserved population has a positive loan experience, Home Credit makes use of a variety of alternative data--including telco and transactional information--to predict their clients repayment abilities.
While Home Credit is currently using various statistical and machine learning methods to make these predictions, theyre challenging Kagglers to help them unlock the full potential of their data. Doing so will ensure that clients capable of repayment are not rejected and that loans are given with a principal, maturity, and repayment calendar that will empower their clients to be successful.",tabular data,featured,Can you predict how capable each applicant is of repaying a loan?,AUC,home-credit-default-risk,auc
House Prices - Advanced Regression Techniques,"Start here if...
You have some experience with R or Python and machine learning basics. This is a perfect competition for data science students who have completed an online course in machine learning and are looking to expand their skill set before trying a featured competition. 
Competition Description
Ask a home buyer to describe their dream house, and they probably wont begin with the height of the basement ceiling or the proximity to an east-west railroad. But this playground competitions dataset proves that much more influences price negotiations than the number of bedrooms or a white-picket fence.
With 79 explanatory variables describing (almost) every aspect of residential homes in Ames, Iowa, this competition challenges you to predict the final price of each home.
Practice Skills
Creative feature engineering 
Advanced regression techniques like random forest and gradient boosting
Acknowledgments
The Ames Housing dataset was compiled by Dean De Cock for use in data science education. Its an incredible alternative for data scientists looking for a modernized and expanded version of the often cited Boston Housing dataset. 
Photo by Tom Thain on Unsplash.",tabular data,getting started,"Predict sales prices and practice feature engineering, RFs, and gradient boosting",rmsle,house-prices-advanced-regression-techniques,percentage_errors
IEEE-CIS Fraud Detection,"Imagine standing at the check-out counter at the grocery store with a long line behind you and the cashier not-so-quietly announces that your card has been declined. In this moment, you probably arent thinking about the data science that determined your fate.
Embarrassed, and certain you have the funds to cover everything needed for an epic nacho party for 50 of your closest friends, you try your card again. Same result. As you step aside and allow the cashier to tend to the next customer, you receive a text message from your bank. Press 1 if you really tried to spend $500 on cheddar cheese.
While perhaps cumbersome (and often embarrassing) in the moment, this fraud prevention system is actually saving consumers millions of dollars per year. Researchers from the IEEE Computational Intelligence Society (IEEE-CIS) want to improve this figure, while also improving the customer experience. With higher accuracy fraud detection, you can get on with your chips without the hassle.
IEEE-CIS works across a variety of AI and machine learning areas, including deep neural networks, fuzzy systems, evolutionary computation, and swarm intelligence. Today theyre partnering with the worlds leading payment service company, Vesta Corporation, seeking the best solutions for fraud prevention industry, and now you are invited to join the challenge.
In this competition, youll benchmark machine learning models on a challenging large-scale dataset. The data comes from Vestas real-world e-commerce transactions and contains a wide range of features from device type to product features. You also have the opportunity to create new features to improve your results.
If successful, youll improve the efficacy of fraudulent transaction alerts for millions of people around the world, helping hundreds of thousands of businesses reduce their fraud loss and increase their revenue. And of course, you will save party people just like you the hassle of false positives.
Acknowledgements:
Vesta Corporation provided the dataset for this competition. Vesta Corporation is the forerunner in guaranteed e-commerce payment solutions. Founded in 1995, Vesta pioneered the process of fully guaranteed card-not-present (CNP) payment transactions for the telecommunications industry. Since then, Vesta has firmly expanded data science and machine learning capabilities across the globe and solidified its position as the leader in guaranteed ecommerce payments. Today, Vesta guarantees more than $18B in transactions annually.
Header Photo by Tim Evans on Unsplash",tabular data,research,Can you detect fraud from customer transactions?,AUC,ieee-cis-fraud-detection,auc
Instacart Market Basket Analysis,"Whether you shop from meticulously planned grocery lists or let whimsy guide your grazing, our unique food rituals define who we are. Instacart, a grocery ordering and delivery app, aims to make it easy to fill your refrigerator and pantry with your personal favorites and staples when you need them. After selecting products through the Instacart app, personal shoppers review your order and do the in-store shopping and delivery for you.
Instacarts data science team plays a big part in providing this delightful shopping experience. Currently they use transactional data to develop models that predict which products a user will buy again, try for the first time, or add to their cart next during a session. Recently, Instacart open sourced this data - see their blog post on 3 Million Instacart Orders, Open Sourced.
In this competition, Instacart is challenging the Kaggle community to use this anonymized data on customer orders over time to predict which previously purchased products will be in a users next order. Theyre not only looking for the best model, Instacarts also looking for machine learning engineers to grow their team.
Winners of this competition will receive both a cash prize and a fast track through the recruiting process. For more information about exciting opportunities at Instacart, check out their careers page here or e-mail their recruiting team directly at ml.jobs@instacart.com.",tabular data,featured,Which products will an Instacart consumer purchase again?,MeanFScore,instacart-market-basket-analysis,f_score_multiclass
Instant Gratification,"Welcome to Instant (well, almost) Gratification!
In 2015, Kaggle introduced Kernels as a resource to competition participants. It was a controversial decision to add a code-sharing tool to a competitive coding space. We thought it was important to make Kaggle more than a place where competitions are solved behind closed digital doors. Since then, Kernels has grown from its infancy--essentially a blinking cursor in a docker container--into its teenage years. We now have more compute, longer runtimes, better datasets, GPUs, and an improved interface.
We have iterated and tested several Kernels-only (KO) competition formats with a true holdout test set, in particular deploying them when we would have otherwise substituted a two-stage competition. However, the experience of submitting to a Kernels-only competition has typically been asynchronous and imperfect; participants wait many days after a competition has concluded for their selected Kernels to be rerun on the holdout test dataset, the leaderboard updated, and the winners announced. This flow causes heartbreak to participants whose Kernels fail on the unseen test set, leaving them with no way to correct tiny errors that spoil months of hard work.
Say Hello to Synchronous KO
Were now pleased to announce general support for a synchronous Kernels-only format. When you submit from a Kernel, Kaggle will run the code against both the public test set and private test set in real time. This small-but-substantial tweak improves the experience for participants, the host, and Kaggle:
With a truly withheld test set, we are practicing proper, rigorous machine learning.
We will be able to offer more varieties of competitions and intend to run many fewer confusing two-stage competitions.
You will be able to see if your code runs successfully on the withheld test set and have the leeway to intervene if it fails.
We will run all submissions against the private data, not just selected ones. Participants will get the complete and familiar public/private scores available in a traditional competition.
The final leaderboard can be released at the end of the competition, without the delay of rerunning Kernels.
This competition is a low-stakes, trial-run introduction to our new synchronous KO implementation. We want to test that the process goes smoothly and gather feedback on your experiences. While it may feel like a normal KO competition, there are complicated new mechanics in play, such as the selection logic of Kernels that are still running when the deadline passes.
Since the competition also presents an authentic machine learning problem, it will also award Kaggle medals and points. Have fun, good luck, and welcome to the world of synchronous Kernels competitions!",tabular data,featured,A synchronous Kernels-only competition,AUC,instant-gratification,auc
Jane Street Market Prediction,"Buy low, sell high. It sounds so easy.
In reality, trading for profit has always been a difficult problem to solve, even more so in todays fast-moving and complex financial markets. Electronic trading allows for thousands of transactions to occur within a fraction of a second, resulting in nearly unlimited opportunities to potentially find and take advantage of price differences in real time.
In a perfectly efficient market, buyers and sellers would have all the agency and information needed to make rational trading decisions. As a result, products would always remain at their fair values and never be undervalued or overpriced. However, financial markets are not perfectly efficient in the real world.
Developing trading strategies to identify and take advantage of inefficiencies is challenging. Even if a strategy is profitable now, it may not be in the future, and market volatility makes it impossible to predict the profitability of any given trade with certainty. As a result, it can be hard to distinguish good luck from having made a good trading decision.
In the first three months of this challenge, you will build your own quantitative trading model to maximize returns using market data from a major global stock exchange. Next, youll test the predictiveness of your models against future market returns and receive feedback on the leaderboard.
Your challenge will be to use the historical data, mathematical tools, and technological tools at your disposal to create a model that gets as close to certainty as possible. You will be presented with a number of potential trading opportunities, which your model must choose whether to accept or reject.
In general, if one is able to generate a highly predictive model which selects the right trades to execute, they would also be playing an important role in sending the market signals that push prices closer to fair values. That is, a better model will mean the market will be more efficient going forward. However, developing good models will be challenging for many reasons, including a very low signal-to-noise ratio, potential redundancy, strong feature correlation, and difficulty of coming up with a proper mathematical formulation.
Jane Street has spent decades developing their own trading models and machine learning solutions to identify profitable opportunities and quickly decide whether to execute trades. These models help Jane Street trade thousands of financial products each day across 200 trading venues around the world.
Admittedly, this challenge far oversimplifies the depth of the quantitative problems Jane Streeters work on daily, and Jane Street is happy with the performance of its existing trading model for this particular question. However, theres nothing like a good puzzle, and this challenge will hopefully serve as a fun introduction to a type of data science problem that a Jane Streeter might tackle on a daily basis. Jane Street looks forward to seeing the new and creative approaches the Kaggle community will take to solve this trading challenge.
This is a Code Competition. Refer to Code Requirements for details.",tabular data,featured,Test your model against future real market data,JaneStreetPnl,jane-street-market-prediction,points
Loan Default Prediction - Imperial College London,"This competition asks you to determine whether a loan will default, as well as the loss incurred if it does default. Unlike traditional finance-based approaches to this problem, where one distinguishes between good or bad counterparties in a binary way, we seek to anticipate and incorporate both the default and the severity of the losses that result. In doing so, we are building a bridge between traditional banking, where we are looking at reducing the consumption of economic capital, to an asset-management perspective, where we optimize on the risk to the financial investor.
This competition is sponsored by researchers at Imperial College London.",tabular data,research,Constructing an optimal portfolio of loans,MAE,loan-default-prediction-imperial-college-london,mae
March Machine Learning Mania 2016,"Update: although the tournament is over, were continuing our analysis under the predictions dataset page.
Back for its third year, March Machine Learning Mania challenges data scientists to predict winners and losers of the mens 2016 NCAA basketball tournament. Youre provided data covering three decades of historical NCAA games and freely encouraged to use other sources of data to gain a winning edge.
In stage one of this two-stage competition, participants will build and test their models against the previous four tournaments. In the second stage, participants will predict the outcome of the 2016 tournament. You dont need to participate in the first stage to enter the second. The first stage exists to incentivize model building and provide a means to score predictions. The real competition is forecasting the 2016 results.
Acknowledgments
SAP is the presenting sponsor of March Machine Learning Mania 2016. Please see About the Sponsor to read more.",tabular data,featured,Predict the 2016 NCAA Basketball Tournament,LogLoss,march-machine-learning-mania-2016,log_loss
Night at Cameo,"Night at Cameo
Illinois Tech ACM designed this event in order to:
Help Illinois Tech students learn about recommender systems
Connect Illinois Tech students with the Cameo team
Explore potential applications of recommender systems to Cameo
Cameo
Learn more about Cameo and their talent on their website.
Challenge
Each team will develop a recommender system to suggest talent that users may be interested in based on talent that the user is known to like.
Get Started
Go to the Kernels tab and fork the starter notebook.
Read Erick Torres article about recommender systems.
Explore the available functions in the redcarpet module.
Use the quick reference to understand the data structures and helper functions.
Photo: Sebastian Ervi on Unsplash",tabular data,inclass,Illinois Institute of Technology ACM + Cameo: Recommend talent that users might like!,map@{k},night-at-cameo,reconstruction
Pneumonia Texture Analysis,"Overview
The goal is to build an algorithm to correctly classify tiles (given their additional metadata like patient age and view) into either opacity or no-opacity groups
Data
The data consists of the metadata in a csv file called _all or _info (for the training and testing groups) as well as the images in a stacked-tiff file where the tile for a given sample can be found by using the slice index.",tabular data,inclass,Performing Texture Analysis on the RSNA Pneumonia Data,auc,pneumonia-texture-analysis,auc
ML 4 Money,"    ,             ?
             -          .
      ?
    -      ,             .
       :
-     -  ,       .
-    MACD, RSI, CCI, DMI, ADL   ",tabular data,inclass,Predict exchange income,rmse,ml-4-money,mse
Oracle Graph ML Contest at Polimi,"In the Machine-Learning track of the contest organized by Politecnico di Milano and Oracle Labs you will have the chance to play with multilabel vertex classification, and with many state-of-the-art vertex embedding algorithms.
Multilabel classification means that we want to assign to each element in a dataset (in our case, to each vertex in a graph) a list of labels, extracted from a well-defined set. For example, you might want to give to a book a list of topics (narrative, sci-fi, fantasy, ). Or you could imagine having a graph where vertices are recipes, and recipes are connected if they have some ingredient in common: in this case, you could predict recipes tags, e.g. spicy or vegetarian.
In this competition, you will work with a list of 24 graphs that represent protein-protein interactions (PPI), and assign to each protein a set of labels that represent its roles.
Truth be told, what we really care about is not the specific domain of the dataset, but the ability to create powerful embeddings for each vertex in these graphs. You will study how to transform the information contained in a vertex (either considering just the graph topology or all its features) to a low-dimensional vector, which can then be fed to a classifier to obtain predictions about this vertex labels.
Vertex embeddings
The study of vertex embeddings has really exploded in the recent past, thanks to the discoveries in the field of Deep Learning and to the abundance of graph data to play with (Google, Facebook, Amazon, ..., you can really make a long list here!). In the Documents page youll get a list of very interesting papers on the topic, that you can use as a starting point for your experiments.
For now, what you want to know is that the starting point of ""modern"" graph embeddings algorithms is called DeepWalk, which creates embeddings by performing random walks on the graph. DeepWalk is available inside Pgx, the graph analysis framework developed by Oracle Labs, and was used to create the first baseline for our predictions.
Your job is to improve this result, either by playing with DeepWalk parameters, by finding a better embedding algorithm or by creating one yourself (dont put limits to your ambitions!).
Important remarks
The main goal of this contest is not to obtain a super amazing prediction accuracy (even though thats also nice to have), but to show that you have a solid grasp of how vertex embeddings work and that you can make good use of them. As such, we are gonna provide you with the classifiers that should be used to make the final predictions. Feel free to fine-tune them a bit to your liking, but dont spend time creating giant ensemble models or things like that, as thats not the skill we are looking for.
The same applies to feature engineering or other things you might be used to do if you have any experience with data-science. Dont spend time hand-crafting features or looking online for more features for your proteins! Get in touch with us if you have doubts or questions :)
Another thing to keep in mind is that you can easily find online pre-made embedding algorithms that will give you great results, and that are basically the state-of-the-art in this field. Some of them are even linked in the papers we have provided! You can use existing algorithms or even existing code, but you should be well aware of what you are doing. Try different algorithms and spend time fine-tuning existing models or making incremental changes to them, and focus on learning what you are doing, instead of simply aiming for the top spot in the leaderboard!
One last important thing! The dataset you are using is freely available online, which means that you can easily cheat in many many different ways, e.g. by training your model on the validation set or by simply copy-pasting the right predictions. You will have to provide us with your code, so that we can check if your results are really farina del tuo sacco or if there is something fishy going on! We decided to use this dataset as its a very common benchmark in the literature, and its easy to see how your results compare with the state-of-the-art.",tabular data,inclass,Contest organized at Polimi with support from Oracle Labs. Create embeddings on a graph and classify its vertices!,meanfscore,oracle-graph-ml-contest-at-polimi,f_score_multiclass
Kharagpur Data Analytics Group,"There are various attributes given for pricing of a house. Help us to build an accurate model to predict the sale for a house.
Note: try to submit your output by private kernel.",tabular data,inclass,Predict the outcome and become the part of our team.,rmse,kharagpur-data-analytics-group,mse
Python for Data science ITEA,"Homework
Create a better model to predict likes on Instagram than your counterparts and get more points.
Grades
1 place - 15 points
2 place - 12 points
3 place - 10 points
4 place - 9 points
5 place - 8 points
6 place - 7 points
7 place - 6 points
8 place - 5 points
9 place - 4 points
10 place - 3 points
11 place - 2 points
12 place - 1 point
Best kernel - 10 points",tabular data,inclass,ML Homework (Instagram likes prediction),smape,python-for-data-science-itea,percentage_errors
kaggle18011884,"2020     
 ",tabular data,inclass,kaggle18011884,rmse,kaggle18011884,mse
M5 Forecasting - Accuracy,"Note: This is one of the two complementary competitions that together comprise the M5 forecasting challenge. Can you estimate, as precisely as possible, the point forecasts of the unit sales of various products sold in the USA by Walmart? If you are interested in estimating the uncertainty distribution of the realized values of the same series, be sure to check out its companion competition
How much camping gear will one store sell each month in a year? To the uninitiated, calculating sales at this level may seem as difficult as predicting the weather. Both types of forecasting rely on science and historical data. While a wrong weather forecast may result in you carrying around an umbrella on a sunny day, inaccurate business forecasts could result in actual or opportunity losses. In this competition, in addition to traditional forecasting methods youre also challenged to use machine learning to improve forecast accuracy.
The Makridakis Open Forecasting Center (MOFC) at the University of Nicosia conducts cutting-edge forecasting research and provides business forecast training. It helps companies achieve accurate predictions, estimate the levels of uncertainty, avoiding costly mistakes, and apply best forecasting practices. The MOFC is well known for its Makridakis Competitions, the first of which ran in the 1980s.
In this competition, the fifth iteration, you will use hierarchical sales data from Walmart, the worlds largest company by revenue, to forecast daily sales for the next 28 days. The data, covers stores in three US States (California, Texas, and Wisconsin) and includes item level, department, product categories, and store details. In addition, it has explanatory variables such as price, promotions, day of the week, and special events. Together, this robust dataset can be used to improve forecasting accuracy.
If successful, your work will continue to advance the theory and practice of forecasting. The methods used can be applied in various business areas, such as setting up appropriate inventory or service levels. Through its business support and training, the MOFC will help distribute the tools and knowledge so others can achieve more accurate and better calibrated forecasts, reduce waste and be able to appreciate uncertainty and its risk implications.
Acknowledgements
Additional thanks go to other partner organizations and prize sponsors, National Technical University of Athens (NTUA), INSEAD, Google, Uber and IIF.",tabular data,featured,Estimate the unit sales of Walmart retail goods,M5_WRMSSE,m5-forecasting-accuracy,percentage_errors
M5 Forecasting - Uncertainty,"Note: This is one of the two complementary competitions that together comprise the M5 forecasting challenge. Can you estimate, as precisely as possible, the uncertainty distribution of the unit sales of various products sold in the USA by Walmart? This specific competition is the first of its kind, opening up new directions for both academic research and how uncertainty could be assessed and used in organizations. If you are interested in providing point (accuracy) forecasts for the same series, be sure to check out its companion competition.
How much camping gear will one store sell each month in a year? To the uninitiated, calculating sales at this level may seem as difficult as predicting the weather. Both types of forecasting rely on science and historical data. While a wrong weather forecast may result in you carrying around an umbrella on a sunny day, inaccurate business forecasts could result in actual or opportunity losses. In this competition, in addition to traditional forecasting methods youre also challenged to use machine learning to improve forecast accuracy.
The Makridakis Open Forecasting Center (MOFC) at the University of Nicosia conducts cutting-edge forecasting research and provides business forecast training. It helps companies achieve accurate predictions, estimate the levels of uncertainty, avoiding costly mistakes, and apply best forecasting practices. The MOFC is well known for its Makridakis Competitions, the first of which ran in the 1980s.
In this competition, the fifth iteration, you will use hierarchical sales data from Walmart, the worlds largest company by revenue, to forecast daily sales for the next 28 days and to make uncertainty estimates for these forecasts. The data, covers stores in three US States (California, Texas, and Wisconsin) and includes item level, department, product categories, and store details. In addition, it has explanatory variables such as price, promotions, day of the week, and special events. Together, this robust dataset can be used to improve forecasting accuracy.
If successful, your work will continue to advance the theory and practice of forecasting. The methods used can be applied in various business areas, such as setting up appropriate inventory or service levels. Through its business support and training, the MOFC will help distribute the tools and knowledge so others can achieve more accurate and better calibrated forecasts, reduce waste and be able to appreciate uncertainty and its risk implications.
Acknowledgements
Additional thanks go to other partner organizations and prize sponsors, National Technical University of Athens (NTUA), INSEAD, Google, Uber and IIF.",tabular data,featured, Estimate the uncertainty distribution of Walmart unit sales.  ,WeightedRowwisePinballLoss,m5-forecasting-uncertainty,custom_loss
NMLO Contest 3 - Regression,"In this competition, you will perform regression using any type of ML model. You will be trying to predict the number of COVID-19 cases in a U.S. county given demographic data. You can use deep learning frameworks and traditional ML libraries (like scikit-learn) for this competition.
Demographic data was acquired from United States Department of Agriculture Economic Research Service.
COVID-19 cases were scraped from Google on 7/18/2020 and 7/19/2020.
You can use any library for this competition.",tabular data,inclass,TJML National Machine Learning Open Contest #3,rmse,nmlo-contest-3-regression,mse
KNIT_HACKS,"Welcome to the knit-hack competition. This competition is mainly for Students of Kamla Nehru Institute of Technology,Sultanpur. However you can participate in the competition but would not be eligible for prizes.",tabular data,inclass,This competion is open for all. But only knit students are eligible to win the cash prize.,macrofscore,knit_hacks,f_score_multiclass
DL for exploration geophysics,"Welcome to our third Kaggle competition!
In this competition, you will create your Deep neural network to predict traveltimes of seismic waves.",tabular data,inclass,Competition #1,rmse,dl-for-exploration-geophysics,mse
WSDM - KKBox's Churn Prediction Challenge,"The 11th ACM International Conference on Web Search and Data Mining (WSDM 2018) is challenging you to build an algorithm that predicts whether a subscription user will churn using a donated dataset from KKBOX. WSDM (pronounced ""wisdom"") is one of the the premier conferences on web inspired research involving search and data mining. Theyre committed to publishing original, high quality papers and presentations, with an emphasis on practical but principled novel models.
For a subscription business, accurately predicting churn is critical to long-term success. Even slight variations in churn can drastically affect profits.
KKBOX is Asias leading music streaming service, holding the worlds most comprehensive Asia-Pop music library with over 30 million tracks. They offer a generous, unlimited version of their service to millions of people, supported by advertising and paid subscriptions. This delicate model is dependent on accurately predicting churn of their paid users.
In this competition youre tasked to build an algorithm that predicts whether a user will churn after their subscription expires. Currently, the company uses survival analysis techniques to determine the residual membership life time for each subscriber. By adopting different methods, KKBOX anticipates theyll discover new insights to why users leave so they can be proactive in keeping users dancing.
Winners will present their findings at the WSDM conference February 6-8, 2018 in Los Angeles, CA. For more information on the conference, click here.",tabular data,research,Can you predict when subscribers will churn?,LogLoss,wsdm-kkboxs-churn-prediction-challenge,log_loss
WSDM - KKBox's Music Recommendation Challenge,"The 11th ACM International Conference on Web Search and Data Mining (WSDM 2018) is challenging you to build a better music recommendation system using a donated dataset from KKBOX. WSDM (pronounced ""wisdom"") is one of the the premier conferences on web inspired research involving search and data mining. Theyre committed to publishing original, high quality papers and presentations, with an emphasis on practical but principled novel models.
Not many years ago, it was inconceivable that the same person would listen to the Beatles, Vivaldi, and Lady Gaga on their morning commute. But, the glory days of Radio DJs have passed, and musical gatekeepers have been replaced with personalizing algorithms and unlimited streaming services.
While the publics now listening to all kinds of music, algorithms still struggle in key areas. Without enough historical data, how would an algorithm know if listeners will like a new song or a new artist? And, how would it know what songs to recommend brand new users?
WSDM has challenged the Kaggle ML community to help solve these problems and build a better music recommendation system. The dataset is from KKBOX, Asias leading music streaming service, holding the worlds most comprehensive Asia-Pop music library with over 30 million tracks. They currently use a collaborative filtering based algorithm with matrix factorization and word embedding in their recommendation system but believe new techniques could lead to better results.
Winners will present their findings at the conference February 6-8, 2018 in Los Angeles, CA. For more information on the conference, click here, and dont forget to check out the other KKBox/WSDM competition: KKBox Music Churn Prediction Challenge",tabular data,research,Can you build the best music recommendation system?,AUC,wsdm-kkboxs-music-recommendation-challenge,auc
Liberty Mutual Group: Property Inspection Prediction,"A Fortune 100 company, Liberty Mutual Insurance has provided a wide range of insurance products and services designed to meet their customers ever-changing needs for over 100 years.
To ensure that Liberty Mutuals portfolio of home insurance policies aligns with their business goals, many newly insured properties receive a home inspection. These inspections review the condition of key attributes of the property, including things like the foundation, roof, windows and siding. The results of an inspection help Liberty Mutual determine if the property is one they want to insure.
In this challenge, your task is to predict a transformed count of hazards or pre-existing damages using a dataset of property information. This will enable Liberty Mutual to more accurately identify high risk homes that require additional examination to confirm their insurability.
Liberty Mutual is interested in hiring predictive modelers like you to work on one of many growing analytics teams within our company. As a member of Liberty Mutuals advanced analytics community, you will have the opportunity to apply sophisticated, cutting-edge techniques, similar to those used in this competition, to large data sets in departments such as Actuarial, Product, Claims, Marketing, Distribution, Human Resources, and Finance. Click to view available positions.
Because we seek to tap innovation both inside and outside the company, certain eligible Liberty Mutual employees are encouraged to participate in this challenge for development purposes. Refer to the competition rules for the full details.",tabular data,featured,Quantify property hazards before time of inspection,NormalizedGini,liberty-mutual-group:-property-inspection-prediction,auc
Mechanisms of Action (MoA) Prediction,"The Connectivity Map, a project within the Broad Institute of MIT and Harvard, the Laboratory for Innovation Science at Harvard (LISH), and the NIH Common Funds Library of Integrated Network-Based Cellular Signatures (LINCS), present this challenge with the goal of advancing drug development through improvements to MoA prediction algorithms.
What is the Mechanism of Action (MoA) of a drug? And why is it important?
In the past, scientists derived drugs from natural products or were inspired by traditional remedies. Very common drugs, such as paracetamol, known in the US as acetaminophen, were put into clinical use decades before the biological mechanisms driving their pharmacological activities were understood. Today, with the advent of more powerful technologies, drug discovery has changed from the serendipitous approaches of the past to a more targeted model based on an understanding of the underlying biological mechanism of a disease. In this new framework, scientists seek to identify a protein target associated with a disease and develop a molecule that can modulate that protein target. As a shorthand to describe the biological activity of a given molecule, scientists assign a label referred to as mechanism-of-action or MoA for short.
How do we determine the MoAs of a new drug?
One approach is to treat a sample of human cells with the drug and then analyze the cellular responses with algorithms that search for similarity to known patterns in large genomic databases, such as libraries of gene expression or cell viability patterns of drugs with known MoAs.
In this competition, you will have access to a unique dataset that combines gene expression and cell viability data. The data is based on a new technology that measures simultaneously (within the same samples) human cells responses to drugs in a pool of 100 different cell types (thus solving the problem of identifying ex-ante, which cell types are better suited for a given drug). In addition, you will have access to MoA annotations for more than 5,000 drugs in this dataset.
As is customary, the dataset has been split into testing and training subsets. Hence, your task is to use the training dataset to develop an algorithm that automatically labels each case in the test set as one or more MoA classes. Note that since drugs can have multiple MoA annotations, the task is formally a multi-label classification problem.
How to evaluate the accuracy of a solution?
Based on the MoA annotations, the accuracy of solutions will be evaluated on the average value of the logarithmic loss function applied to each drug-MoA annotation pair.
If successful, youll help to develop an algorithm to predict a compounds MoA given its cellular signature, thus helping scientists advance the drug discovery process.
This is a Code Competition. Refer to Code Requirements for details.",tabular data,research,Can you improve the algorithm that classifies drugs based on their biological activity?,MeanColumnwiseLogLoss,mechanisms-of-action-(moa)-prediction,multiclass_log_loss
University of Liverpool - Ion Switching,"Think you can use your data science skills to make big predictions at a submicroscopic level?
Many diseases, including cancer, are believed to have a contributing factor in common. Ion channels are pore-forming proteins present in animals and plants. They encode learning and memory, help fight infections, enable pain signals, and stimulate muscle contraction. If scientists could better study ion channels, which may be possible with the aid of machine learning, it could have a far-reaching impact.
When ion channels open, they pass electric currents. Existing methods of detecting these state changes are slow and laborious. Humans must supervise the analysis, which imparts considerable bias, in addition to being tedious. These difficulties limit the volume of ion channel current analysis that can be used in research. Scientists hope that technology could enable rapid automatic detection of ion channel current events in raw data.
The University of Liverpools Institute of Ageing and Chronic Disease is working to advance ion channel research. Their team of scientists have asked for your help. In this competition, youll use ion channel data to better model automatic identification methods. If successful, youll be able to detect individual ion channel events in noisy raw signals. The data is simulated and injected with real world noise to emulate what scientists observe in laboratory experiments.
Technology to analyze electrical data in cells has not changed significantly over the past 20 years. If we better understand ion channel activity, the research could impact many areas related to cell health and migration. From human diseases to how climate change affects plants, faster detection of ion channels could greatly accelerate solutions to major world problems.
Acknowledgements:
This would not be possible without the help of the Biotechnology and Biological Sciences Research Council (BBSRC).",tabular data,research,Identify the number of channels open at each time point,MacroFScore,university-of-liverpool-ion-switching,f_score_multiclass
Mercari Price Suggestion Challenge,"It can be hard to know how much somethings really worth. Small details can mean big differences in pricing. For example, one of these sweaters cost $335 and the other cost $9.99. Can you guess which ones which?
Product pricing gets even harder at scale, considering just how many products are sold online. Clothing has strong seasonal pricing trends and is heavily influenced by brand names, while electronics have fluctuating prices based on product specs.
Mercari, Japans biggest community-powered shopping app, knows this problem deeply. Theyd like to offer pricing suggestions to sellers, but this is tough because their sellers are enabled to put just about anything, or any bundle of things, on Mercaris marketplace.
In this competition, Mercaris challenging you to build an algorithm that automatically suggests the right product prices. Youll be provided user-inputted text descriptions of their products, including details like product category name, brand name, and item condition.
Note that, because of the public nature of this data, this competition is a Kernels Only competition. In the second stage of the challenge, files will only be available through Kernels and you will not be able to modify your approach in response to new data. Read more details in the data tab and Kernels FAQ page.",tabular data,featured,Can you automatically suggest product prices to online sellers?,RMSLE,mercari-price-suggestion-challenge,percentage_errors
Mercedes-Benz Greener Manufacturing,"Since the first automobile, the Benz Patent Motor Car in 1886, Mercedes-Benz has stood for important automotive innovations. These include, for example, the passenger safety cell with crumple zone, the airbag and intelligent assistance systems. Mercedes-Benz applies for nearly 2000 patents per year, making the brand the European leader among premium car makers. Daimlers Mercedes-Benz cars are leaders in the premium car industry. With a huge selection of features and options, customers can choose the customized Mercedes-Benz of their dreams. .
To ensure the safety and reliability of each and every unique car configuration before they hit the road, Daimlers engineers have developed a robust testing system. But, optimizing the speed of their testing system for so many possible feature combinations is complex and time-consuming without a powerful algorithmic approach. As one of the worlds biggest manufacturers of premium cars, safety and efficiency are paramount on Daimlers production lines.
In this competition, Daimler is challenging Kagglers to tackle the curse of dimensionality and reduce the time that cars spend on the test bench. Competitors will work with a dataset representing different permutations of Mercedes-Benz car features to predict the time it takes to pass testing. Winning algorithms will contribute to speedier testing, resulting in lower carbon dioxide emissions without reducing Daimlers standards.",tabular data,featured,Can you cut the time a Mercedes-Benz spends on the test bench?,R2Score,mercedes-benz-greener-manufacturing,r2_score
Microsoft Malware Prediction,"The malware industry continues to be a well-organized, well-funded market dedicated to evading traditional security measures. Once a computer is infected by malware, criminals can hurt consumers and enterprises in many ways.
With more than one billion enterprise and consumer customers, Microsoft takes this problem very seriously and is deeply invested in improving security.
As one part of their overall strategy for doing so, Microsoft is challenging the data science community to develop techniques to predict if a machine will soon be hit with malware. As with their previous, Malware Challenge (2015), Microsoft is providing Kagglers with an unprecedented malware dataset to encourage open-source progress on effective techniques for predicting malware occurrences.
Can you help protect more than one billion machines from damage BEFORE it happens?
Acknowledgements
This competition is hosted by Microsoft, Windows Defender ATP Research, Northeastern University College of Computer and Information Science, and Georgia Tech Institute for Information Security & Privacy.
Microsoft contacts
Rob McCann (Robert.McCann@microsoft.com)
Christian Seifert (chriseif@microsoft.com)
Susan Higgs (Susan.Higgs@microsoft.com)
Matt Duncan (Matthew.Duncan@microsoft.com)
Northeastern University contact
Mansour Ahmadi (m.ahmadi@northeastern.edu)
Georgia Tech contacts
Brendan Saltaformaggio (brendan@ece.gatech.edu)
Taesoo Kim (taesoo@gatech.edu)",tabular data,research,Can you predict if a machine will soon be hit with malware?,AUC,microsoft-malware-prediction,auc
New York City Taxi Fare Prediction,"In this playground competition, hosted in partnership with Google Cloud and Coursera, you are tasked with predicting the fare amount (inclusive of tolls) for a taxi ride in New York City given the pickup and dropoff locations. While you can get a basic estimate based on just the distance between the two points, this will result in an RMSE of $5-$8, depending on the model used (see the starter code for an example of this approach in Kernels). Your challenge is to do better than this using Machine Learning techniques!
To learn how to handle large datasets with ease and solve this problem using TensorFlow, consider taking the Machine Learning with TensorFlow on Google Cloud Platform specialization on Coursera -- the taxi fare problem is one of several real-world problems that are used as case studies in the series of courses. To make this easier, head to Coursera.org/NEXTextended to claim this specialization for free for the first month!",tabular data,playground,Can you predict a rider's taxi fare?,RMSE,new-york-city-taxi-fare-prediction,mse
NFL Big Data Bowl,"The running back takes the handoff he breaks a tacklespins and breaks free! One man to beat! Past the 50-yard-line! To the 40! The 30! He! Could! Go! All! The! Way!
But will he?
American football is a complex sport. From the 22 players on the field to specific characteristics that ebb and flow throughout the game, it can be challenging to quantify the value of specific plays and actions within a play. Fundamentally, the goal of football is for the offense to run (rush) or throw (pass) the ball to gain yards, moving towards, then across, the opposing teams side of the field in order to score. And the goal of the defense is to prevent the offensive team from scoring.
In the National Football League (NFL), roughly a third of teams offensive yardage comes from run plays.. Ball carriers are generally assigned the most credit for these plays, but their teammates (by way of blocking), coach (by way of play call), and the opposing defense also play a critical role. Traditional metrics such as yards per carry or total rushing yards can be flawed; in this competition, the NFL aims to provide better context into what contributes to a successful run play.
As an armchair quarterback watching the game, you may think you can predict the result of a play when a ball carrier takes the handoff - but what does the data say? In this competition, you will develop a model to predict how many yards a team will gain on given rushing plays as they happen. Youll be provided game, play, and player-level data, including the position and speed of players as provided in the NFLs Next Gen Stats data. And the best part - you can see how your model performs from your living room, as the leaderboard will be updated week after week on the current seasons game data as it plays out.
Deeper insight into rushing plays will help teams, media, and fans better understand the skill of players and the strategies of coaches. It will also assist the NFL and its teams evaluate the ball carrier, his teammates, his coach, and the opposing defense, in order to make adjustments as necessary.
Additionally, the winning model will be provided to the NFLs Next Gen Stats group to potentially share with teams. You could help the NFL Network generate models to use during games, or for pre-game/post-game breakdowns.",tabular data,featured,How many yards will an NFL player gain after receiving a handoff?,CRPS,nfl-big-data-bowl,custom_loss
Nomad2018 Predicting Transparent Conductors,"Innovative materials design is needed to tackle some of the most important health, environmental, energy, social, and economic challenges of this century. In particular, improving the properties of materials that are intrinsically connected to the generation and utilization of energy is crucial if we are to mitigate environmental damage due to a growing global demand. Transparent conductors are an important class of compounds that are both electrically conductive and have a low absorption in the visible range, which are typically competing properties. A combination of both of these characteristics is key for the operation of a variety of technological devices such as photovoltaic cells, light-emitting diodes for flat-panel displays, transistors, sensors, touch screens, and lasers. However, only a small number of compounds are currently known to display both transparency and conductivity suitable enough to be used as transparent conducting materials.
Aluminum Al, gallium Ga, indium In sesquioxides are some of the most promising transparent conductors because of a combination of both large bandgap energies, which leads to optical transparency over the visible range, and high conductivities. These materials are also chemically stable and relatively inexpensive to produce. Alloying of these binary compounds in ternary or quaternary mixtures could enable the design of a new material at a specific composition with improved properties over what is current possible. These alloys are described by the formula (AlxGayInz)2NO3N; where x, y, and z can vary but are limited by the constraint x+y+z = 1. The total number of atoms in the",tabular data,research,Predict the key properties of novel transparent semiconductors,MCRMSLE,nomad2018-predicting-transparent-conductors,percentage_errors_multiclass
New York City Taxi Trip Duration,"In this competition, Kaggle is challenging you to build a model that predicts the total ride duration of taxi trips in New York City. Your primary dataset is one released by the NYC Taxi and Limousine Commission, which includes pickup time, geo-coordinates, number of passengers, and several other variables.
Longtime Kagglers will recognize that this competition objective is similar to the ECML/PKDD trip time challenge we hosted in 2015. But, this challenge comes with a twist. Instead of awarding prizes to the top finishers on the leaderboard, this playground competition was created to reward collaboration and collective learning.
We are encouraging you (with cash prizes!) to publish additional training data that other participants can use for their predictions. We also have designated bi-weekly and final prizes to reward authors of kernels that are particularly insightful or valuable to the community.",tabular data,playground,Share code and data to improve ride time predictions,RMSLE,new-york-city-taxi-trip-duration,percentage_errors
PMR3508 - Tarefa 1 - 3508 Adult Dataset,"Introduo
Prezados Alunos,
Bem vindos  primeira etapa da primeira tarefa de PMR3508. Esta tarefa est dividada em duas etapas. A primeira consiste em participar desta competio fechada, em um ambiente seguro para que vocs se familiarizem com o Kaggle, com o Python e com suas ferramentas. A segunda parte, consiste em se inscrever na competio [Costa Rican Household Poverty Level Prediction](https://www.kaggle.com/c/costa-rican-household-poverty-prediction) e fazer o mesmo procedimento que ser feito aqui.
Procedimento
O procedimento para execuo desta tarefa  bem simples. Espera-se que cada aluno faa um notebook do Jupyter que contenha ao menos duas seces: Uma de explorao de dados, em que ele observa como so distribudas as variveis do dataset, quais valores elas assumem etc; a outra dedicada  avaliao do impacto da seleo de variveis e feature engineering e seleo do parmetro K, do algoritmo K-Nearest Neighbors(KNN), sobre a acurcia do seu classificador.
A submisso das suas previses para o set de teste  opcional, mas encorajada, j que, para incentivar um pouco de competitividade saudvel, o monitor promete uma barra de chocolate Lindt ( de 125 gramas) ao estudante vencedor da leaderboard que tiver seu resultado comprovado no kernel publicado. Em caso de veganismo, diabetes, intolerncia a lactose ou outro imprevisto que prejudique o usufruto o prmio pelo vencedor, um substituto de valor similar ser negociado em tempo hbil.
ATENO: A nota deste exerccio ser dada com base no KERNEL SUMETIDO. Caso o aluno submeta um resultado para a leaderboard, mas no fornea o kernel, ELE NO S NO ESTAR COMPETINDO PELO CHOCOLATE, COMO RECEBER NOTA ZERO NESTA ATIVIDADE
Instrues
Na tab ""Data"" vocs podero encontrar os arquivos contendo os dados e suas respectivas descries.
Aps a execuo de sua tarefa, salve seu notebook (em formato .ipynb) e clique na aba ""Kernels"". L, clique em ""new kernel"" e siga as instrues para submeter seu kernel na competio. Em seguida, proceda de igual forma para a segunda parte deste exerccio.
Agradecimentos especiais ao Repositrio UCI por ter providenciado este dataset ao domnio pblico.
Dicas
1) A documentao do pandas  sua amiga
2) O comando pd.describe()  muito til para ter uma noo dos seus dados
3) A biblioteca ScikitLearn possui vrios tutoriais interessantes em seu site ( http://scikit-learn.org/stable/index.html )
4) Lembrem-se de fazer validao cruzada de seus resultados!!!!",tabular data,inclass,Introdução ao Kaggle e ao algoritmo k-means,categorizationaccuracy,pmr3508-tarefa-1-3508-adult-dataset,categorization_accuracy
Influencers in Social Networks,"Data Science London and the UK Windows Azure Users Group in partnership with Microsoft and Peerindex, announce the Influencers in Social Networks competition as part of The Big Data Hackathon.  
The dataset, provided by Peerindex, comprises a standard, pair-wise preference learning task. Each datapoint describes two individuals, A and B. For each person, 11 p
The binary label represents a human judgement about which one of the two individuals is more influential. A label 1 means A is more influential than B. 0 means B is more influential than A. The goal of the challenge is to train a machine learning model which, for pairs of individuals, predicts the human judgement on who is more influential with high accuracy. Labels for the dataset have been collected by PeerIndex using an application similar to the one described in this post.
A python script computing a sample benchmark solution is available here: 
Competition begins: Saturday, Apr 13, 1pm BST (12 noon UTC)
This competition awards 25% the ranking points of a standard competition, but does not count towards tiers. ",tabular data,featured,Predict which people are influential in a social network,AUC,influencers-in-social-networks,auc
StumbleUpon Evergreen Classification Challenge,"StumbleUpon is a user-curated web content discovery engine that recommends relevant, high quality pages and media to its users, based on their interests. While some pages we recommend, such as news articles or seasonal recipes, are only relevant for a short period of time, others maintain a timeless quality and can be recommended to users long after they are discovered. In other words, pages can either be classified as ""ephemeral"" or ""evergreen"". The ratings we get from our community give us strong signals that a page may no longer be relevant - but what if we could make this distinction ahead of time? A high quality prediction of ""ephemeral"" or ""evergreen"" would greatly improve a recommendation system like ours.
Many people know evergreen content when they see it, but can an algorithm make the same determination without human intuition? Your mission is to build a classifier which will evaluate a large set of URLs and label them as either evergreen or ephemeral. Can you out-class(ify) StumbleUpon? As an added incentive to the prize, a strong performance in this competition may lead to a career-launching internship at one of the best places to work in San Francisco.",tabular data,featured,Build a classifier to categorize webpages as evergreen or non-evergreen,AUC,stumbleupon-evergreen-classification-challenge,auc
Restaurant Revenue Prediction,"With over 1,200 quick service restaurants across the globe, TFI is the company behind some of the worlds most well-known brands: Burger King, Sbarro, Popeyes, Usta Donerci, and Arbys. They employ over 20,000 people in Europe and Asia and make significant daily investments in developing new restaurant sites.
Right now, deciding when and where to open new restaurants is largely a subjective process based on the personal judgement and experience of development teams. This subjective data is difficult to accurately extrapolate across geographies and cultures. 
New restaurant sites take large investments of time and capital to get up and running. When the wrong location for a restaurant brand is chosen, the site closes within 18 months and operating losses are incurred. 
Finding a mathematical model to increase the effectiveness of investments in new restaurant sites would allow TFI to invest more in other important business areas, like sustainability, innovation, and training for new employees. Using demographic, real estate, and commercial data, this competition challenges you to predict the annual restaurant sales of 100,000 regional locations.
TFI would love to hire an expert Kaggler like you to head up their growing data science team in Istanbul or Shanghai. Youd be tackling problems like the one featured in this competition on a global scale. See the job description here >>",tabular data,featured,Predict annual restaurant sales based on objective measurements,RMSE,restaurant-revenue-prediction,mse
West Nile Virus Prediction,"West Nile virus is most commonly spread to humans through infected mosquitos. Around 20% of people who become infected with the virus develop symptoms ranging from a persistent fever, to serious neurological illnesses that can result in death.
In 2002, the first human cases of West Nile virus were reported in Chicago. By 2004 the City of Chicago and the Chicago Department of Public Health (CDPH) had established a comprehensive surveillance and control program that is still in effect today.
Every week from late spring through the fall, mosquitos in traps across the city are tested for the virus. The results of these tests influence when and where the city will spray airborne pesticides to control adult mosquito populations.
Given weather, location, testing, and spraying data, this competition asks you to predict when and where different species of mosquitos will test positive for West Nile virus. A more accurate method of predicting outbreaks of West Nile virus in mosquitos will help the City of Chicago and CPHD more efficiently and effectively allocate resources towards preventing transmission of this potentially deadly virus. 
Weve jump-started your analysis with some visualizations and starter code in R and Python on Kaggle Scripts. No data download or local environment setup needed!
Acknowledgements 
This competition is sponsored by the Robert Wood Johnson Foundation. Data is provided by the Chicago Department of Public Health.",tabular data,featured,Predict West Nile virus in mosquitos across the city of Chicago,AUC,west-nile-virus-prediction,auc
Predicting Red Hat Business Value,"Like most companies, Red Hat is able to gather a great deal of information over time about the behavior of individuals who interact with them. Theyre in search of better methods of using this behavioral data to predict which individuals they should approachand even when and how to approach them.
In this competition, Kagglers are challenged to create a classification algorithm that accurately identifies which customers have the most potential business value for Red Hat based on their characteristics and activities.
With an improved prediction model in place, Red Hat will be able to more efficiently prioritize resources to generate more business and better serve their customers.",tabular data,featured,Classify customer potential,AUC,predicting-red-hat-business-value,auc
TalkingData Mobile User Demographics,"Nothing is more comforting than being greeted by your favorite drink just as you walk through the door of the corner caf. While a thoughtful barista knows you take a macchiato every Wednesday morning at 8:15, its much more difficult in a digital space for your preferred brands to personalize your experience.
TalkingData, Chinas largest third-party mobile data platform, understands that everyday choices and behaviors paint a picture of who we are and what we value. Currently, TalkingData is seeking to leverage behavioral data from more than 70% of the 500 million mobile devices active daily in China to help its clients better understand and interact with their audiences.
In this competition, Kagglers are challenged to build a model predicting users demographic characteristics based on their app usage, geolocation, and mobile device properties. Doing so will help millions of developers and brand advertisers around the world pursue data-driven marketing efforts which are relevant to their users and catered to their preferences.
Acknowledgements",tabular data,featured,Get to know millions of mobile device users,MulticlassLoss,talkingdata-mobile-user-demographics,multiclass_log_loss
Santa's Uncertain Bags,"All was well in Santas workshop. The gifts were made, the route was planned, the naughty and nice list complete. Santa thought this would finally be the year he didnt need Kaggles help with his combinatorial conundrums. At last, the Claus family could take the elves and reindeer on that well deserved vacation to the South Pole.
Then, with just days until the big night, Santa received an email from a panicked database admin elf. Attached was a server log with the six least jolly words a jolly old St. Nick could read:
ALTER TABLE Gifts
DROP COLUMN Weight
One of the North Pole elf interns had mistakenly deleted the weights for all of the inventory in the workshop! Santa didnt have a backup (remember, this is a guy who makes a list and checks it twice) and, without knowing each presents weight, he didnt know how he would safely pack his many gift bags. Gifts were already on their way to the sleigh packing facility and there wasnt time to re-weigh all the presents. It was once again necessary to summon the holiday talents of Kaggles elite.
Can you help Santa fill his multiple bags with sets of uncertain gifts? Save the season by turning Santas uncertain probabilities into presents for good little boys and girls.",tabular data,playground,"♫ Bells are ringing, children singing, all is merry and bright. Santa's elves made a big mistake, now he needs your help tonight ♫",SantaWeightedBins,santas-uncertain-bags,points
Tap30 Challenge,"Estimating supply and demand is one of the most important problems for online transportation companies. Many of these companies large scale business strategies depend on being able to accurately predict supply and demand at any point in time. For instance knowing that during certain times in the day the number of requests for rides exceeds the number of available drivers might lead the compnay to encourage drivers to work more during those times by providing incentives.
In this challenge you are provided with data indicating the number of requests for rides per hour in different areas of Tehran spanning a period of several weeks. Note that the data provided to you for this challenge contains the actual number of requests observed by Tap30 and you are facing a problem with real world data.",tabular data,inclass,Online Taxi Demand Prediction,rmse,tap30-challenge,mse
Santa Gift Matching Challenge,"Tis the night before Christmas
year: two thousand seventeen.
Santas grown grouchy,
borderline mean.
What used to be simple for Old St. Nick,
is now too puzzling, its making him sick!
See, Santa always knew, deep down in his gut,
what toy each kid wantedno ifs, ands, or buts.
But fierce population growth, more twins, and toy innovation,
has left too complex a problem, in dire need of optimization.
Dont worry, Mr. Santa, said an Elf named McMaggle,
I have a solution! Have you heard of Kaggle?
As she explained Kaggle in-depth, Santas doubt began turning,
he became a believer in the magic of...machine learning.
So, Santas team needs YOU more than ever this year,
to solve this painful problem and save Christmas cheer.
The Challenge
In this playground competition, youre challenged to build a toy matching algorithm that maximizes happiness by pairing kids with toys they want. In the dataset, each kid has 10 preferences for their gift (from 1000) and Santa has 1000 preferred kids for every gift available. What makes this extra difficult is that 0.4% of the kids are twins, and by their parents request, require the same gift.",tabular data,featured,Down through the chimney with lots of toys...,SantaResident,santa-gift-matching-challenge,significance
SQL Saturday Madrid ML Challenge,"This is the home page of the competition. You dont need a subtitle here. The competition sub-title will appear above. This is where you introduce the problem. You can upload images using the ""select files"" widget on the left in the competition wizard. Upload an image, refresh the page, copy its URL, then insert within the wizards editor.
If you are copy-pasting from another application, like Word or your browser, try to make sure the html formatting is clean. You can view a pages html using the button at the top right of the editors toolbar.
This is a subtitle
To format pages, stick to the following conventions:
Paragraphs should go in p tags
Code should go in pre tags
Subtitles should go in h2 tags
You can display equations using LaTeX enclosed in escaped brackets. For example, this:
\[ \epsilon = \sqrt{\frac{1}{n} \sum_{i=1}^n (\log(p_i + 1) - \log(a_i+1))^2 } \]
is created by this:
\[
\epsilon = \sqrt{\frac{1}{n} \sum_{i=1}^n (\log(p_i + 1) - \log(a_i+1))^2 }
\]
Acknowledgements
We thank Professor Plum, Ph.D. for providing this dataset.",tabular data,inclass,Demuestra lo que sabes de Machine Learning con PASS España!,meanfscore,sql-saturday-madrid-ml-challenge,f_score_multiclass
Pycon Korea 2018 - Tutorial,"Pycon Korea 2018    Tutorial .
https://www.pycon.kr/2018/program/tutorial/13
  Kaggle   Forest Cover Type.

 Multiclass Binary Classification .




Random forests? Cover trees? Not so fast, computer nerds. Were talking about the real thing.
In this competition you are asked to predict the forest cover type (the predominant kind of tree cover) from cartographic variables. The actual forest cover type for a given 30 x 30 meter cell was determined from US Forest Service (USFS) Region 2 Resource Information System data. Independent variables were then derived from data obtained from the US Geological Survey and USFS. The data is in raw form and contains binary columns of data for qualitative independent variables such as wilderness areas and soil type.
This study area includes four wilderness areas located in the Roosevelt National Forest of northern Colorado. These areas represent forests with minimal human-caused disturbances, so that existing forest cover types are more a result of ecological processes rather than forest management practices.
Acknowledgements
This dataset was provided by Jock A. Blackard and Colorado State University. We also thank the UCI machine learning repository for hosting the dataset. If you use the problem in publication, please cite:
Bache, K. & Lichman, M. (2013). UCI Machine Learning Repository. Irvine, CA: University of California, School of Information and Computer Science",tabular data,inclass,Pycon Korea 2018 - 미운 우리 캐글 Tutorial 공간입니다.,logloss,pycon-korea-2018-tutorial,log_loss
INGV - Volcanic Eruption Prediction,"What if scientists could anticipate volcanic eruptions as they predict the weather? While determining rain or shine days in advance is more difficult, weather reports become more accurate on shorter time scales. A similar approach with volcanoes could make a big impact. Just one unforeseen eruption can result in tens of thousands of lives lost. If scientists could reliably predict when a volcano will next erupt, evacuations could be more timely and the damage mitigated.
Currently, scientists often identify time to eruption by surveying volcanic tremors from seismic signals. In some volcanoes, this intensifies as volcanoes awaken and prepare to erupt. Unfortunately, patterns of seismicity are difficult to interpret. In very active volcanoes, current approaches predict eruptions some minutes in advance, but they usually fail at longer-term predictions.
Enter Italys Istituto Nazionale di Geofisica e Vulcanologia (INGV), with its focus on geophysics and volcanology. The INGVs main objective is to contribute to the understanding of the Earths system while mitigating the associated risks. Tasked with the 24-hour monitoring of seismicity and active volcano activity across the country, the INGV seeks to find the earliest detectable precursors that provide information about the timing of future volcanic eruptions.
In this competition, using your data science skills, youll predict when a volcanos next eruption will occur. Youll analyze a large geophysical dataset collected by sensors deployed on active volcanoes. If successful, your algorithms will identify signatures in seismic waveforms that characterize the development of an eruption.
With enough notice, areas around a volcano can be safely evacuated prior to their destruction. Seismic activity is a good indicator of an impending eruption, but earlier precursors must be identified to improve longer-term predictability. The impact of your participation could be felt worldwide with tens of thousands of lives saved by more predictable volcanic ruptures and earlier evacuations.",tabular data,playground,Discover hidden precursors in geophysical data to help emergency response,MAE,ingv-volcanic-eruption-prediction,mae
Prudential Life Insurance Assessment,"Picture this. You are a data scientist in a start-up culture with the potential to have a very large impact on the business. Oh, and you are backed up by a company with 140 years business experience.
Curious? Great! You are the kind of person we are looking for.
Prudential, one of the largest issuers of life insurance in the USA, is hiring passionate data scientists to join a newly-formed Data Science group solving complex challenges and identifying opportunities. The results have been impressive so far but we want more. 
The Challenge
In a one-click shopping world with on-demand
everything, the life insurance application process is antiquated. Customers provide extensive information to identify risk classification and eligibility, including scheduling medical exams, a process that takes an average of 30 days.
The result? People are turned off. Thats why only 40% of U.S. households own individual life insurance. Prudential wants to make it quicker and less labor intensive for new and existing customers to get a quote while maintaining privacy boundaries.
By developing a predictive model that accurately classifies risk using a more automated approach, you can greatly impact public perception of the industry.
The results will help Prudential better understand the predictive power of the data points in the existing assessment, enabling us to significantly streamline the process.",tabular data,featured,Can you make buying life insurance easier?,QuadraticWeightedKappa,prudential-life-insurance-assessment,correlation
PUBG Finish Placement Prediction (Kernels Only),"So, where we droppin boys and girls?
Battle Royale-style video games have taken the world by storm. 100 players are dropped onto an island empty-handed and must explore, scavenge, and eliminate other players until only one is left standing, all while the play zone continues to shrink.
PlayerUnknowns BattleGrounds (PUBG) has enjoyed massive popularity. With over 50 million copies sold, its the fifth best selling game of all time, and has millions of active monthly players.
The team at PUBG has made official game data available for the public to explore and scavenge outside of ""The Blue Circle."" This competition is not an official or affiliated PUBG site - Kaggle collected data made possible through the PUBG Developer API.
You are given over 65,000 games worth of anonymized player data, split into training and testing sets, and asked to predict final placement from final in-game stats and initial player ratings.
Whats the best strategy to win in PUBG? Should you sit in one spot and hide your way into victory, or do you need to be the top shot? Lets let the data do the talking!",tabular data,playground,Can you predict the battle royale finish of PUBG Players?,MAE,pubg-finish-placement-prediction-(kernels-only),mae
Santa 2019 - Revenge of the Accountants,"Santa was thrilled with the Kaggle community for minimizing his workshop costs! He had heard rumors that Kagglers were adept at cracking holiday challenges, but, wow, even Santa was surprised at this one.
Unfortunately, the North Pole accountants were less pleased. It turns out, the accountants didnt like being one-upped by machine learning experts on the internet.
To complicate matters, theyve decided to allow an additional 1,000 families attend the workshop. And theyve also ""fine tuned"" their accounting formula to try and trip up those fancy solvers some people have at their disposal.
Of course, we know that nothing trips up the Kaggle community! (Well, except for maybe over-fitting. But fortunately, that doesnt apply here!)
So this is a bonus Santa competition for those who want an additional challenge and the opportunity to continue to improve their optimization skills. Since Santa used up all his budget on accounting fees, this is strictly a Playground competition, with the chance to win some coveted Kaggle Swag.
Have fun, and Happy Holidays from the Kaggle Team!
Attribution
Banner/Listing Photo by Helloquence on Unsplash",tabular data,playground,Oh what fun it is to revise . . .,SantaWorkshopSchedule2019Revenge,santa-2019-revenge-of-the-accountants,custom_loss
Santa's Workshop Tour 2019,"Hammers ring, are you listenin
In the shop, toys are glistenin
Should they see the sights?
There might be a fight
Walkin round the Workshop Wonderland
Families said, they want to see it
Santa said, hed guarantee it
They pick a date
But they may have to wait
Walkin round the Workshop Wonderland
We told Santa that he was a madman
He just wants to make sure they all smile
Hell say Are you flexible?, Theyll say Yeah man,
But can you help us make it worth our while?
Give them food, or sweater
the more they wait, the gifts get better
Please help us rank
Or well break the bank!
Walkin round the Workshop Wonderland
Santa has exciting news! For 100 days before Christmas, he opened up tours to his workshop. Because demand was so strong, and because Santa wanted to make things as fair as possible, he let each of the 5,000 families that will visit the workshop choose a list of dates theyd like to attend the workshop.
Now that all the families have sent Santa their preferences, hes realized its impossible for everyone to get their top picks, so hes decided to provide extra perks for families that dont get their preferences. In addition, Santas accounting department has told him that, depending on how families are scheduled, there may be some unexpected and hefty costs incurred.
Santa needs the help of the Kaggle community to optimize which day each family is assigned to attend the workshop in order to minimize any extra expenses that would cut into next years toy budget! Can you help Santa out?
Attribution
Banner/Listing Photo by Nathan Lemon on Unsplash
Description Photo by Markus Spiske on Unsplash",tabular data,featured,"In the notebook we can build a model, and pretend that it will optimize...",SantaWorkshopSchedule2019,santas-workshop-tour-2019,custom_loss
Santander Customer Satisfaction,"From frontline support teams to C-suites, customer satisfaction is a key measure of success. Unhappy customers dont stick around. Whats more, unhappy customers rarely voice their dissatisfaction before leaving.
Santander Bank is asking Kagglers to help them identify dissatisfied customers early in their relationship. Doing so would allow Santander to take proactive steps to improve a customers happiness before its too late.
In this competition, youll work with hundreds of anonymized features to predict if a customer is satisfied or dissatisfied with their banking experience.",tabular data,featured,Which customers are happy customers?,AUC,santander-customer-satisfaction,auc
Santander Customer Transaction Prediction,"At Santander our mission is to help people and businesses prosper. We are always looking for ways to help our customers understand their financial health and identify which products and services might help them achieve their monetary goals.
Our data science team is continually challenging our machine learning algorithms, working with the global data science community to make sure we can more accurately identify new ways to solve our most common challenge, binary classification problems such as: is a customer satisfied? Will a customer buy this product? Can a customer pay this loan?
In this challenge, we invite Kagglers to help us identify which customers will make a specific transaction in the future, irrespective of the amount of money transacted. The data provided for this competition has the same structure as the real data we have available to solve this problem.",tabular data,featured,Can you identify who will make a transaction?,AUC,santander-customer-transaction-prediction,auc
Santander Value Prediction Challenge,"According to Epsilon research, 80% of customers are more likely to do business with you if you provide personalized service. Banking is no exception.
The digitalization of everyday lives means that customers expect services to be delivered in a personalized and timely manner and often before theyve even realized they need the service. In their 3rd Kaggle competition, Santander Group aims to go a step beyond recognizing that there is a need to provide a customer a financial service and intends to determine the amount or value of the customers transaction. This means anticipating customer needs in a more concrete, but also simple and personal way. With so many choices for financial services, this need is greater now than ever before.
In this competition, Santander Group is asking Kagglers to help them identify the value of transactions for each potential customer. This is a first step that Santander needs to nail in order to personalize their services at scale.",tabular data,featured,Predict the value of transactions for potential customers.,RMSLE,santander-value-prediction-challenge,percentage_errors
Sberbank Russian Housing Market,"Housing costs demand a significant investment from both consumers and developers. And when it comes to planning a budgetwhether personal or corporatethe last thing anyone needs is uncertainty about one of their biggets expenses. Sberbank, Russias oldest and largest bank, helps their customers by making predictions about realty prices so renters, developers, and lenders are more confident when they sign a lease or purchase a building.
Although the housing market is relatively stable in Russia, the countrys volatile economy makes forecasting prices as a function of apartment characteristics a unique challenge. Complex interactions between housing features such as number of bedrooms and location are enough to make pricing predictions complicated. Adding an unstable economy to the mix means Sberbank and their customers need more than simple regression models in their arsenal.
In this competition, Sberbank is challenging Kagglers to develop algorithms which use a broad spectrum of features to predict realty prices. Competitors will rely on a rich dataset that includes housing data and macroeconomic patterns. An accurate forecasting model will allow Sberbank to provide more certainty to their customers in an uncertain economy.",tabular data,featured,Can you predict realty price fluctuations in Russia’s volatile economy?,RMSLE,sberbank-russian-housing-market,percentage_errors
American Epilepsy Society Seizure Prediction Challenge,"Seizure forecasting systems hold promise for improving the quality of life for patients with epilepsy.
Epilepsy afflicts nearly 1% of the worlds population, and is characterized by the occurrence of spontaneous seizures. For many patients, anticonvulsant medications can be given at sufficiently high doses to prevent seizures, but patients frequently suffer side effects. For 20-40% of patients with epilepsy, medications are not effective -- and even after surgical removal of epilepsy-causing brain tissue, many patients continue to experience spontaneous seizures. Despite the fact that seizures occur infrequently, patients with epilepsy experience persistent anxiety due to the possibility of a seizure occurring.
Seizure forecasting systems have the potential to help patients with epilepsy lead more normal lives. In order for EEG-based seizure forecasting systems to work effectively, computational algorithms must reliably identify periods of increased probability of seizure occurrence. If these seizure-permissive brain states can be identified, devices designed to warn patients of impeding seizures would be possible. Patients could avoid potentially dangerous activities like driving or swimming, and medications could be administered only when needed to prevent impending seizures, reducing overall side effects.
There is emerging evidence that the temporal dynamics of brain activity can be classified into 4 states: Interictal (between seizures, or baseline), Preictal (prior to seizure), Ictal (seizure), and Post-ictal (after seizures). Seizure forecasting requires the ability to reliably identify a preictal state that can be differentiated from the interictal, ictal, and postictal state. The primary challenge in seizure forecasting is differentiating between the preictal and interictal states. The goal of the competition is to demonstrate the existence and accurate classification of the preictal brain state in dogs and humans with naturally occurring epilepsy.
The Competition
Intracranial EEG was recorded from dogs with naturally occurring epilepsy using an ambulatory monitoring system. EEG was sampled from 16 electrodes at 400 Hz, and recorded voltages were referenced to the group average. These are long duration recordings, spanning multiple months up to a year and recording up to a hundred seizures in some dogs.
In addition, datasets from patients with epilepsy undergoing intracranial EEG monitoring to identify a region of brain that can be resected to prevent future seizures are included in the contest. These datasets have varying numbers of electrodes and are sampled at 5000 Hz, with recorded voltages referenced to an electrode outside the brain. The challenge is to distinguish between ten minute long data clips covering an hour prior to a seizure, and ten minute iEEG clips of interictal activity. Seizures are known to cluster, or occur in groups. Patients who typically have seizure clusters receive little benefit from forecasting follow-on seizures. For this contest only lead seizures, defined here as seizures occurring four hours or more after another seizure, are included in the training and testing data sets. In order to avoid any potential contamination between interictal, preictal, and post-ictal EEG signals interictal segments in the canine training and test data were restricted to be at least one week before or after any seizure. In the human data, where the entire monitoring session may last less than one week, interictal data segments were restricted to be at least four hours before or after any seizure. Interictal data segments were chosen at random within these restrictions for both canine and human subjects.
Participants are invited to visit the NIH-sponsored International Epilepsy Electrophysiology portal (http://ieeg.org) to review and download annotated interictal and preictal data from other patients and animal subjects. Using ieeg.org data for additional algorithm training is permitted.
Acknowledgements
This competition is sponsored by the National Institutes of Health (NINDS), the Epilepsy Foundation, and the American Epilepsy Society.
  References
Howbert JJ, Patterson EE, Stead SM, Brinkmann B, Vasoli V, Crepeau D, Vite CH, Sturges B, Ruedebusch V, Mavoori J, Leyde K, Sheffield WD, Litt B, Worrell GA (2014) Forecasting seizures in dogs with naturally occurring epilepsy. PLoS One 9(1):e81920.
Cook MJ, OBrien TJ, Berkovic SF, Murphy M, Morokoff A, Fabinyi G, DSouza W, Yerra R, Archer J, Litewka L, Hosking S, Lightfoot P, Ruedebusch V, Sheffield WD, Snyder D, Leyde K, Himes D (2013) Prediction of seizure likelihood with a long-term, implanted seizure advisory system in patients with drug-resistant epilepsy: a first-in-man study. LANCET NEUROL 12:563-571.
Park Y, Luo L, Parhi KK, Netoff T (2011) Seizure prediction with spectral power of EEG using cost-sensitive support vector machines. Epilepsia 52:1761-1770.
Davis KA, Sturges BK, Vite CH, Ruedebusch V, Worrell G, Gardner AB, Leyde K, Sheffield WD, Litt B (2011) A novel implanted device to wirelessly record and analyze continuous intracranial canine EEG. Epilepsy Res 96:116-122.
Andrzejak RG, Chicharro D, Elger CE, Mormann F (2009) Seizure prediction: Any better than chance? Clin Neurophysiol.
Snyder DE, Echauz J, Grimes DB, Litt B (2008) The statistics of a practical seizure warning system. J Neural Eng 5: 392401.
Mormann F, Andrzejak RG, Elger CE, Lehnertz K (2007) Seizure prediction: the long and winding road. Brain 130: 314333.
Haut S, Shinnar S, Moshe SL, ODell C, Legatt AD. (1999) The association between seizure clustering and status epilepticus in patients with intractable complex partial seizures. Epilepsia 40:18321834.",tabular data,research,Predict seizures in intracranial EEG recordings,AUC,american-epilepsy-society-seizure-prediction-challenge,auc
San Francisco Crime Classification,"From 1934 to 1963, San Francisco was infamous for housing some of the worlds most notorious criminals on the inescapable island of Alcatraz.
Today, the city is known more for its tech scene than its criminal past. But, with rising wealth inequality, housing shortages, and a proliferation of expensive digital toys riding BART to work, there is no scarcity of crime in the city by the bay.
From Sunset to SOMA, and Marina to Excelsior, this competitions dataset provides nearly 12 years of crime reports from across all of San Franciscos neighborhoods. Given time and location, you must predict the category of crime that occurred.
Were also encouraging you to explore the dataset visually. What can we learn about the city through visualizations like this Top Crimes Map? The top most up-voted scripts from this competition will receive official Kaggle swag as prizes. 
Acknowledgements
Kaggle is hosting this competition for the machine learning community to use for fun and practice. This dataset is brought to you by SF OpenData, the central clearinghouse for data published by the City and County of San Francisco.",tabular data,playground,Predict the category of crimes that occurred in the city by the bay ,MulticlassLoss,san-francisco-crime-classification,multiclass_log_loss
Shelter Animal Outcomes,"Every year, approximately 7.6 million companion animals end up in US shelters. Many animals are given up as unwanted by their owners, while others are picked up after getting lost or taken out of cruelty situations. Many of these animals find forever families to take them home, but just as many are not so lucky. 2.7 million dogs and cats are euthanized in the US every year.
Using a dataset of intake information including breed, color, sex, and age from the Austin Animal Center, were asking Kagglers to predict the outcome for each animal.
We also believe this dataset can help us understand trends in animal outcomes. These insights could help shelters focus their energy on specific animals who need a little extra help finding a new home. We encourage you to publish your insights on Scripts so they are publicly accessible.
Acknowledgements
Kaggle is hosting this competition for the machine learning community to use for data science practice and social good. The dataset is brought to you by Austin Animal Center. Shelter animal statistics were taken from the ASPCA.
Glamour shots of Kaggles shelter pets are pictured above. From left to right: Shelby, Bailey, Hazel, Daisy, and Yeti.",tabular data,playground,Help improve outcomes for shelter animals,MulticlassLoss,shelter-animal-outcomes,multiclass_log_loss
Springleaf Marketing Response,"Springleaf puts the humanity back into lending by offering their customers personal and auto loans that help them take control of their lives and their finances. Direct mail is one important way Springleafs team can connect with customers whom may be in need of a loan.
Direct offers provide huge value to customers who need them, and are a fundamental part of Springleafs marketing strategy. In order to improve their targeted efforts, Springleaf must be sure they are focusing on the customers who are likely to respond and be good candidates for their services.
Using a large set of anonymized features, Springleaf is asking you to predict which customers will respond to a direct mail offer. You are challenged to construct new meta-variables and employ feature-selection methods to approach this dauntingly wide dataset.",tabular data,featured,Determine whether to send a direct mail piece to a customer ,AUC,springleaf-marketing-response,auc
Tabular Playground Series - Feb 2021,"Kaggle competitions are incredibly fun and rewarding, but they can also be intimidating for people who are relatively new in their data science journey. In the past, weve launched many Playground competitions that are more approachable than our Featured competitions and thus, more beginner-friendly.
In order to have a more consistent offering of these competitions for our community, were trying a new experiment in 2021. Well be launching month-long tabular Playground competitions on the 1st of every month and continue the experiment as long as theres sufficient interest and participation.
The goal of these competitions is to provide a fun, and approachable for anyone, tabular dataset. These competitions will be great for people looking for something in between the Titanic Getting Started competition and a Featured competition. If youre an established competitions master or grandmaster, these probably wont be much of a challenge for you. We encourage you to avoid saturating the leaderboard.
For each monthly competition, well be offering Kaggle Merchandise for the top three teams. And finally, because we want these competitions to be more about learning, were limiting team sizes to 3 individuals.
The dataset is used for this competition is synthetic, but based on a real dataset and generated using a CTGAN. The original dataset deals with predicting the amount of an insurance claim. Although the features are anonymized, they have properties relating to real-world features.
Good luck and have fun!
Getting Started
Check out this Starter Notebook which walks you through how to make your very first submission!
For more ideas on how to improve your score, check out the Intro to Machine Learning and Intermediate Machine Learning courses on Kaggle Learn.",tabular data,playground,Practice your ML skills on this approachable dataset!,RMSE,tabular-playground-series-feb-2021,mse
Tabular Playground Series - Jan 2021,"Kaggle competitions are incredibly fun and rewarding, but they can also be intimidating for people who are relatively new in their data science journey. In the past, weve launched many Playground competitions that are more approachable than our Featured competitions, and thus more beginner-friendly.
In order to have a more consistent offering of these competitions for our community, were trying a new experiment in 2021. Well be launching a month-long tabular Playground competition on the 1st of every month, and continue the experiment as long as theres sufficient interest and participation.
The goal of these competitions is to provide a fun, but less challenging, tabular dataset. These competitions will be great for people looking for something in between the Titanic Getting Started competition and a Featured competition. If youre an established competitions master or grandmaster, these probably wont be much of a challenge for you. We encourage you to avoid saturating the leaderboard.
For each monthly competition, well be offering Kaggle Merchandise for the top three teams. And finally, because we want these competitions to be more about learning, were limiting team sizes to 3 individuals.
Good luck and have fun!
Getting Started
Check out this Starter Notebook which walks you through how to make your very first submission!
For more ideas on how to improve your score, check out the Intro to Machine Learning and Intermediate Machine Learning courses on Kaggle Learn.",tabular data,playground,Practice your ML regression skills on this approachable dataset!,RMSE,tabular-playground-series-jan-2021,mse
TalkingData AdTracking Fraud Detection Challenge,"Fraud risk is everywhere, but for companies that advertise online, click fraud can happen at an overwhelming volume, resulting in misleading click data and wasted money. Ad channels can drive up costs by simply clicking on the ad at a large scale. With over 1 billion smart mobile devices in active use every month, China is the largest
mobile market in the world and therefore suffers from huge volumes of fradulent traffic.
TalkingData, Chinas largest independent big data service platform, covers over 70% of active mobile devices nationwide. They handle 3 billion clicks per day, of which 90% are potentially fraudulent. Their current approach to prevent click fraud for app developers is to measure the journey of a users click across their portfolio, and flag IP addresses who produce lots of clicks, but never end up installing apps. With this information, theyve built an IP blacklist and device blacklist.
While successful, they want to always be one step ahead of fraudsters and have turned to the Kaggle community for help in further developing their solution. In their 2nd competition with Kaggle, youre challenged to build an algorithm that predicts whether a user will download an app after clicking a mobile app ad. To support your modeling, they have provided a generous dataset covering approximately 200 million clicks over 4 days!",tabular data,featured,Can you detect fraudulent click traffic for mobile app ads?,AUC,talkingdata-adtracking-fraud-detection-challenge,auc
Traveling Santa 2018 - Prime Paths,"Rudolph the red-nosed reindeer
Had some very tired hooves
But he had a job to finish
Could he do it with the shortest moves?
All of the other reindeer
Used to laugh and mock his code
They always said poor Rudolph
Couldnt handle the workload
Then one foggy Christmas Eve
Santa came to say
I see youve taken number theory
Please make this night a bit less dreary?
Then how the reindeer loved him
and each enrolled in an AI degree
Rudolph the red-nosed reindeer
We get to go to bed early!
Rudolph has always believed in working smarter, not harder. And what better way to earn the respect of Comet and Blitzen than showing the initiative to improve Santas annual route for delivering toys on Christmas Eve?
This year, Rudolph believes he can motivate the overworked Reindeer team by wisely choosing the order in which they visit the houses on Santas list. The houses in prime cities always leave carrots for the Reindeers alongside the usual cookies and milk. These carrots are just the sustenance the Reindeers need to keep pace. In fact, Rudolph has found that if the Reindeer team doesnt originate from a prime city exactly every 10th step, it takes the 10% longer than it normally would to make their next destination!
Can you help Rudolph solve the Traveling Santa problem subject to his carrot constraint? His team--and Santa--are counting on you!
Attributions:
Reindeer Photo: Norman Tsui
Stocking Photo: Wesley Tingey",tabular data,featured,"But does your code recall, the most efficient route of all?",TravelingSanta2,traveling-santa-2018-prime-paths,custom_loss
Porto Seguro’s Safe Driver Prediction,"Nothing ruins the thrill of buying a brand new car more quickly than seeing your new insurance bill. The sting’s even more painful when you know you’re a good driver. It doesn’t seem fair that you have to pay so much if you’ve been cautious on the road for years. Porto Seguro, one of Brazil’s largest auto and homeowner insurance companies, completely agrees. Inaccuracies in car insurance company’s claim predictions raise the cost of insurance for good drivers and reduce the price for bad ones. In this competition, you’re challenged to build a model that predicts the probability that a driver will initiate an auto insurance claim in the next year. While Porto Seguro has used machine learning for the past 20 years, they’re looking to Kaggle’s machine learning community to explore new, more powerful methods. A more accurate prediction will allow them to further tailor their prices, and hopefully make auto insurance coverage more accessible to more drivers.",tabular data,featured,Predict if a driver will file an insurance claim next year.,NormalizedGini,porto-seguro’s-safe-driver-prediction,auc
LANL Earthquake Prediction,"Forecasting earthquakes is one of the most important problems in Earth science because of their devastating consequences. Current scientific studies related to earthquake forecasting focus on three key points: when the event will occur, where it will occur, and how large it will be. In this competition, you will address when the earthquake will take place. Specifically, you’ll predict the time remaining before laboratory earthquakes occur from real-time seismic data. If this challenge is solved and the physics are ultimately shown to scale from the laboratory to the field, researchers will have the potential to improve earthquake hazard assessments that could save lives and billions of dollars in infrastructure. This challenge is hosted by Los Alamos National Laboratory which enhances national security by ensuring the safety of the U.S. nuclear stockpile, developing technologies to reduce threats from weapons of mass destruction, and solving problems related to energy, environment, infrastructure, health, and global security concerns.",tabular data,research,Can you predict upcoming laboratory earthquakes?,MAE,lanl-earthquake-prediction,mae
Zillow Prize: Zillow’s Home Value Prediction (Zestimate),"Zillow’s Zestimate home valuation has shaken up the U.S. real estate industry since first released 11 years ago A home is often the largest and most expensive purchase a person makes in his or her lifetime. Ensuring homeowners have a trusted way to monitor this asset is incredibly important. The Zestimate was created to give consumers as much information as possible about homes and the housing market, marking the first time consumers had access to this type of home value information at no cost. “Zestimates” are estimated home values based on 7.5 million statistical and machine learning models that analyze hundreds of data points on each property. And, by continually improving the median margin of error (from 14% at the onset to 5% today), Zillow has since become established as one of the largest, most trusted marketplaces for real estate information in the U.S. and a leading example of impactful machine learning. Zillow Prize, a competition with a one million dollar grand prize, is challenging the data science community to help push the accuracy of the Zestimate even further. Winning algorithms stand to impact the home values of 110M homes across the U.S. In this million-dollar competition, participants will develop an algorithm that makes predictions about the future sale prices of homes. The contest is structured into two rounds, the qualifying round which opens May 24, 2017 and the private round for the 100 top qualifying teams that opens on Feb 1st, 2018. In the qualifying round, you’ll be building a model to improve the Zestimate residual error. In the final round, you’ll build a home valuation algorithm from the ground up, using external data sources to help engineer new features that give your model an edge over the competition. Because real estate transaction data is public information, there will be a three-month sales tracking period after each competition round closes where your predictions will be evaluated against the actual sale prices of the homes. The final leaderboard won’t be revealed until the close of the sales tracking period.",tabular data,featured,Can you improve the algorithm that changed the world of real estate?,ZillowMAE,zillow-prize:-zillow’s-home-value-prediction-(zestimate),mae
2019 Data Science Bowl,"Illuminate Learning. Ignite Possibilities. Uncover new insights in early childhood education and how media can support learning outcomes. Participate in our fifth annual Data Science Bowl, presented by Booz Allen Hamilton and Kaggle. PBS KIDS, a trusted name in early childhood education for decades, aims to gain insights into how media can help children learn important skills for success in school and life. In this challenge, you’ll use anonymous gameplay data, including knowledge of videos watched and games played, from the PBS KIDS Measure Up! app, a game-based learning tool developed as a part of the CPB-PBS Ready To Learn Initiative with funding from the U.S. Department of Education. Competitors will be challenged to predict scores on in-game assessments and create an algorithm that will lead to better-designed games and improved learning outcomes. Your solutions will aid in discovering important relationships between engagement with high-quality educational media and learning processes. Data Science Bowl is the world’s largest data science competition focused on social good. Each year, this competition gives Kagglers a chance to use their passion to change the world. Over the last four years, more than 50,000+ Kagglers have submitted over 114,000+ submissions, to improve everything from lung cancer and heart disease detection to ocean health. For more information on the Data Science Bowl, please visit DataScienceBowl.com Where does the data for the competition come from? The data used in this competition is anonymous, tabular data of interactions with the PBS KIDS Measure Up! app. Select data, such as a user’s in-app assessment score or their path through the game, is collected by the PBS KIDS Measure Up! app, a game-based learning tool. PBS KIDS is committed to creating a safe and secure environment that family members of all ages can enjoy. The PBS KIDS Measure Up! app does not collect any personally identifying information, such as name or location. All of the data used in the competition is anonymous. To view the full PBS KIDS privacy policy, please visit: pbskids.org/privacy. No one will be able to download the entire data set and the participants do not have access to any personally identifiable information about individual users. The Data Science Bowl and the use of data for this year’s competition has been reviewed to ensure that it meets requirements of applicable child privacy regulations by PRIVO, a leading global industry expert in children’s online privacy. What is the PBS KIDS Measure Up! app? In the PBS KIDS Measure Up! app, children ages 3 to 5 learn early STEM concepts focused on length, width, capacity, and weight while going on an adventure through Treetop City, Magma Peak, and Crystal Caves. Joined by their favorite PBS KIDS characters, children can also collect rewards and unlock digital toys as they play. To learn more about PBS KIDS Measure Up!, please click here. PBS KIDS and the PBS KIDS Logo are registered trademarks of PBS. Used with permission. The contents of PBS KIDS Measure Up! were developed under a grant from the Department of Education. However, those contents do not necessarily represent the policy of the Department of Education, and you should not assume endorsement by the Federal Government. The app is funded by a Ready To Learn grant (PR/AWARD No. U295A150003, CFDA No. 84.295A) provided by the Department of Education to the Corporation for Public Broadcasting.",tabular data,featured,Uncover the factors to help measure how young children learn,QuadraticWeightedKappa,data-science-bowl-2019,correlation
Lyft Motion Prediction for Autonomous Vehicles,"Autonomous vehicles (AVs) are expected to dramatically redefine the future of transportation. However, there are still significant engineering challenges to be solved before one can fully realize the benefits of self-driving cars. One such challenge is building models that reliably predict the movement of traffic agents around the AV, such as cars, cyclists, and pedestrians. The ridesharing company Lyft started Level 5 to take on the self-driving challenge and build a full self-driving system (they’re hiring!). Their previous competition tasked participants with identifying 3D objects, an important step prior to detecting their movement. Now, they’re challenging you to predict the motion of these traffic agents. In this competition, you’ll apply your data science skills to build motion prediction models for self-driving vehicles. Youll have access to the largest Prediction Dataset ever released to train and test your models. Your knowledge of machine learning will then be required to predict how cars, cyclists,and pedestrians move in the AVs environment. Lyft’s mission is to improve people’s lives with the world’s best transportation. They believe in a future where self-driving cars make transportation safer, environment-friendly and more accessible for everyone. Their goal is to accelerate development across the industry by sharing data with researchers. As a result of your participation, you can have a hand in propelling the industry forward and helping people around the world benefit from self-driving cars sooner.",tabular data,featured,Build motion prediction models for self-driving vehicles ,PostProcessorKernel,lyft-motion-prediction-autonomous-vehicles,custom_loss
Google Cloud & NCAA® ML Competition 2018-Women's,"Google Cloud and NCAA® have teamed up to bring you this year’s version of the Kaggle machine learning competition. Another year, another chance to anticipate the upsets, call the probabilities, and put your bracketology skills to the leaderboard test. Kagglers will join the millions of fans who attempt to forecast the outcomes of March Madness® during this years NCAA Division I Men’s and Women’s Basketball Championships. But unlike most fans, you will pick your bracket using a combination of NCAA’s historical data and your computing power, while the ground truth unfolds on national television. In the first stage of the competition, Kagglers will rely on results of past tournaments to build and test models. We encourage you to post any useful external data as a dataset. In the second stage, competitors will forecast outcomes of all possible match-ups in the 2018 NCAA Division I Men’s and Women’s Basketball Championships. You dont need to participate in the first stage to enter the second. The first stage exists to incentivize model building and provide a means to score predictions. The real competition is forecasting the 2018 results.",tabular data,featured,Apply machine learning to NCAA® March Madness®,LogLoss,womens-machine-learning-competition-2018,log_loss
Google Cloud & NCAA® ML Competition 2019-Women's,"As a result of the continued collaboration between Google Cloud and the NCAA®, the sixth annual Kaggle-backed March Madness competition is underway! Another year, another chance to anticipate the upsets, call the probabilities, and put your bracketology skills to the leaderboard test. Kagglers will join the millions of fans who attempt to forecast the outcomes of March Madness during this years NCAA Division I Men’s and Women’s Basketball Championships. But unlike most fans, you will pick your bracket using a combination of NCAA’s historical data and your computing power, while the ground truth unfolds on national television.In the first stage of the competition, Kagglers will rely on results of past tournaments to build and test models. We encourage you to post any useful external data as a dataset. In the second stage, competitors will forecast outcomes of all possible matchups in the 2019 NCAA Division I Men’s and Women’s Basketball Championships. You dont need to participate in the first stage to enter the second. The first stage exists to incentivize model building and provide a means to score predictions. The real competition is forecasting the 2019 results. As the official public cloud provider of the NCAA, Google Cloud is proud to provide a competition to help participants strengthen their knowledge of basketball, statistics, data modeling, and cloud technology. As part of its journey to the cloud, the NCAA has migrated 80+ years of historical and play-by-play data, from 90 championships and 24 sports, to Google Cloud Platform (GCP). The NCAA has tapped into decades of historical basketball data using BigQuery, Cloud Spanner, Datalab, Cloud Machine Learning and Cloud Dataflow, to power the analysis of team and player performance. The mission of the NCAA has long been about serving the needs of schools, their teams and students. Google Cloud is proud to support that mission by helping the NCAA use data and machine learning to better engage with its millions of fans, 500,000 student-athletes and more than 19,000 teams. Game on!",tabular data,featured,Apply Machine Learning to NCAA® March Madness®,LogLoss,womens-machine-learning-competition-2019,log_loss
Bike Sharing Demand for Education(),"BikeSharingDSG
test
BikeSharingDemand2011120126train, 20127~12testtest50%public,private

2120

submit
120()

Base_submission
notebook
DR_submission
AutoML
Best_submission
",tabular data,inclass,DSG内での学習用にBikeSharingDemandデータを利用して作成したコンペです,rmsle,bike-sharing-demand-for-education(),percentage_errors
ACM Summer'19 Inclass-1,"This is the first inclass kaggle competition for ACM ML SMP Summer 19
The goal is to learn how to go about a kaggle competition and make the process seem less daunting. Please go through the starter kernel. Feel free to explore other kernels for ideas and hacks that could improve your score on the leader board
By the end of this competition we hope you learn the importance of cross validation schemes, experiment with feature engineering, feature scaling, exploratory data analysis, hyper parameter tuning, feature encoding and maybe even other machine learning models like xgboost and lightgbm etc. The goal is to learn as much as possible",tabular data,inclass,Here is a kaggle competition using a Graduate Admission dataset,rmse,acm-summer19-inclass-1,mse
compass-canada,"You like Molsons, eh?",tabular data,inclass,"if you know, you know...eh",rmse,compass-canada,mse
Temperature Forecasting,"Context
Japan has four distinct seasons.
About 1,300 AMeDAS (Automated Meteorological Data Acquisition System) stations automatically record weather data such as temperature, sunshine, and windspeed.
Temperature is a seasonal time-series data.
It contains a seasonality component and there may be a trend caused by climate change.
Content
This dataset is about minimum temperature forecasting.
This dataset is acquired from Japan Meteorological Agency.
The data contains 4,018 consecutive instances (days), the first 3,653 instances are for training, the last 365 instances are for evaluation.
This competition is inspired from Daily Minimum Temperatures in Melbourne dataset in Kaggle.
Training Data and Test Data
The training data is 10-year (3,653-day) consecutive data, that includes minimum air temperature, maximum air temperature, sunshine, rainfall, and mean windspeed.
The test data is the subsequent 365 days.
You can also use maximum air temperature, sunshine, rainfall, and mean windspeed in training data, but test data does not contain them.
Reference
Japan Meteorological Agency
Paul Brabban: Daily Minimum Temperatures in Melbourne, Kaggle Datasets",tabular data,inclass,The 10th 1056Lab Data Analytics Competition,rmse,temperature-forecasting,mse
SMEMI309 - Final evaluation challenge 2020,"This Kaggle challenge is your last evaluation in the SMEMI309 class (Computational Statistical Methods, by Prof. Jean Martinet) in 2020. The challenge will be held between Nov 19 (Thursday) and December 10 (Thursday) 20.00. Time is in UTC.
NEW: teams of 3 students max allowed (not mandatory).
Context
The objective is to classify data generated by a Spiking Neural Network (SNN). The general approach is described in this paper.
Warning
This challenge contains real, unexplored, freshly generated experimental data from the on-going research of our lab. Students may or may not find useful results during the challenge. Inversely, the data may or may not be too easy to process.
This makes the challenge realistic, exciting, and challenging. This also makes your work particularly important since your results might be useful for a research group. The main point is to find relevant answers, by any means. You are free to use any technique learnt in the class or elsewhere.
Experimental settings and data description
A webcam is plugged to an event-camera simulator, and shown several stimuli, in the form of object motion in four directions (up, down, left, right). The event-camera simulator is a piece of software written to stimulate an event-camera.
The event data is fed to a one-layer SNN with 10 output neurons, and the output activity is recorded for a fixed short duration. One sample output vector is made of the output neuron spike code during this duration, i.e. 10 integer values. We will study three different neural encoding.
Rate coding (output neuron spike count - onsc): The code is made of the total count of spikes during the stimulation.
Temporal latency coding (output neuron spike temporal::first - onstf): The code, also called time-to-first-spike, is made of the latency of the first spike, i.e. the time between stimulus and the first spike.
Temporal rank-order coding (output neuron spike temporal::order - onsto): The code is made of the order of arrival of first spikes.
The datasets are obtained by presenting an object (a hand) moving in translation before the webcam with a fixed speed, in one of the four directions. This has been repeated 80 times for each class. Original 320 samples has been rotated by 90, 180, and 270 degrees for augmentation. Moreover, this data augmentation removes any bias related to the video acquisition, since each sequence now belongs in all four classes. The resulting 1280 event sequences are fed to the network, and we obtain 320 10-D vectors, that are split into train / test sets. Classes are balanced in all sets. Class numbers are 0 for up, 1 for down, 2 for left, and 3 for right.
There are three datasets:
BEFORE_TRAINING (BT) obtained on a randomly initialised SNN,
AFTER_TRAINING with 2 classes (AT2) obtained with the same network (random init.), after a short STDP-based unsupervised training using just two classes (for you to find out!), and
AFTER_TRAINING with 4 classes (AT4), obtained with the same network (trained with 2 classes), after a short STDP-based unsupervised training using all four classes.
The purpose of separating AT2 and AT4 datasets is to verify if the network is able to successfully lean 2 classes, and then to successfully learn 4 classes, with continuous adaptation of varying number of classes (continuous learning).
The main dataset is AT4. This means that you need to use AT4_train to train, and AT4_test to generate your submission file. Note that all test datasets are generated using four classes, even AT2.
All three encoding schemes are in single files in the following order:
label, 10 columns for count, 10 columns for temporal::first, 10 columns for temporal::order
BT_train.csv, AT2_train.csv and AT4_train.csv (1280 rows, 31 columns)
label, onsc1, onsc2,  onsc10, onstf1, onstf2,  onstf10, onso1, onso2,  onso10
1,4,4,  8,0.6002,0.6002,  0.5202,6,6,  0
1,0,0,  6,3.01,3.01,  0.6402,-1,-1,  1
0,2,3,  5,0.7602,0.8801999999999999,  0,0,  0
etc.
Importante notes:
""-1"" means that the neuron did not spike
the arbitrary number of digit gives you full precision
but digits were rounded to evaluate rank orders.
BT_test.csv, AT2_test.csv and AT4_test.csv (320 rows, 30 columns because of course, no label!)
onsc1, onsc2,  onsc10, onstf1, onstf2,  onstf10, onso1, onso2,  onso10
3,3,  ,9,0.6402,0.6402,  ,0.6002,4,4,  ,0
3,2,  4,0.8801999999999999,0.8801999999999999,  0.8402,4,4,  ,0
4,3,  4,1.1201999999999999,1.1201999999999999,  1.1201999999999999,1,1,1
etc.
sampleSubmission.csv (320 rows, 2 columns)
id, label
0, 0
1, 0
2, 0
3, 0
4, 0
etc.
Expected work and results
Four parts are expected (the analysis should be done for ALL 3 encoding schemes):
A statistical analysis to determine whether the data in BT_train.csv, AT2_train.csv, and AT4_train.csv significantly differ from one another. Of course, the hypothesis under test is that the data differ.
The main task in this challenge is the training of a model based on AT4_train.csv to correctly classify AT4_test.csv. You are expected to work on a Kaggle notebook and to post frequently your results (2 submissions per day allowed).
Another task (not to be posted) consists in training a model on BT_train.csv to correctly classify BT_test.csv. Do not try too hard to get good results with this dataset, since we expect (and even hope) this result to be bad.
The last task consists in checking the continuous learning capability of the network.
Please give explicit sections names in your report. In your conclusion, explicitly state what neural coding enables the best classification?
Grade
The grade will take into account (subject to modifications):
(4/10 points) a synthetic description of your solution, including the statistical analysis and detailed explanation of your choices, and why you believe they are good (4-pages max pdf) -- PDF not ipynb. The report is to be submitted on LMS
(4/10 points) the last submitted solution for AT4_test
(2/10 points) the final challenge ranking (max for first)
Note that the ranking formula is coef x [ 1 - (rank-1) / (nbTeams-1) ]",tabular data,inclass,Classify SNN output data by all possible means,categorizationaccuracy,smemi309-final-evaluation-challenge-2020,categorization_accuracy
2019S UTS Data Analytics Assignment 3,"This third assignment picks up from your second assignment, where you explored a dataset. Now, you will build a classifier to predict the QuoteConversionFlag of the Insurance Marketing Data. The classification goal is to predict whether the customer will buy the insurance or not (target attribute: QuoteConversionFlag (binary: 0, 1)).
Important: only users with their UTS student ID as a user-name will be eligible for the class prize.",tabular data,inclass,Predict whether the customer will buy the insurance or not !!,auc,2019s-uts-data-analytics-assignment-3,auc
Diabetes Diagnosis,"Context
This dataset is about diabetes that is a lifelong condition that causes a persons blood sugar level to become too high [NHS].
There are 2 main types of diabetes:
type 1 diabetes  where the bodys immune system attacks and destroys the cells that produce insulin
type 2 diabetes  where the body does not produce enough insulin or the bodys cells do not react to insulin
Type 2 diabetes is far more common than type 1. In the UK, around 90% of all adults with diabetes have type 2.
The key figures about diabetes are [IDF]:
1 in 11 adults (20-79 years) have diabetes (463 million people).
1 in 2 adults with diabetes are undiagnosed (232 million people).
1 in 5 people with diabetes are above 65 years old (136 million people).
10% of global health expenditure is spent on diabetes (USD760 billion).
1 in 6 live births (20 million) is affected by hyperglycemia in pregnancy, 84% of which have gestational diabetes.
3 in 4 (79%) of people with diabetes live in low- and middle-income countries.
Over 1.1 million children and adolescents below 20 yeas have type 1 diabetes.
1 in 13 adults (20-79 years) have impaired glucose tolerance (374 million people).
2 in 3 people with diabetes live in urban areas (310.3 million).
Content
Original data came from the Biostatistics program at Vanderbilt. This dataset was downloaded from data.world.
The downloaded dataset contains 390 instances (patients) including 60 instances are diabetes and the remaining 330 are not diabetes.
Any patient without a Hemoglobin A1c was excluded. If their Hemoglobin A1c was 6.5 or greater they were labeled with diabetes = 1. The downloaded dataset contains Glucose but it was deleted from this dataset because it is strongly related to diabetes.
Training Data and Test Data
The training data is randomly chosen 273 (70%) instances.
The test data is the remaining 117 (30%) instances.
Notice
This is a private competition in 1056Lab, a data mining laboratory at Chubu University.
Reference
Diabetes Prediction - data.world
Diabetes - The National Health Service (NHS), UK
IDF DIABETES ATLAS, 9th edition 2019 - International Diabetes Federation",tabular data,inclass,The 17th 1056Lab Data Analytics Competion,auc,diabetes-diagnosis,auc
Used Cars Price Prediction,"Context
What determines the price of used cars?
The value of a car drops right from the moment it is bought and the depreciation continues with each passing year. In fact, in the first year itself, the value of a car decreases by 20 percent of its initial value. The make and model of a car, total kilometers driven, overall condition of the vehicle and various other factors further affect the cars resale value [CarDekho].
Content
The original dataset is in the Kaggle Datasets.
This data contains 6,019 instances (cars) includes 3,205 diesel cars, 2,746 petrol cars, 56 CNG cars, 10 LPG cars, and 2 electric cars (probably) in India.
Training Data and Test Data
The training data is randomly chosen 4,213 (70%) instances. The test data is the remaining 1,806 (30%) instances.
Notice
This is a private competition in 1056Lab, a data mining laboratory at Chubu University.
Reference
Avi Kasliwal: User Cars Price Prediction - Kaggle Datasets
Frequently Asked Questions - CarDekho",tabular data,inclass,The 18th 1056Lab Data Analytics Competition,rmsle,used-cars-price-prediction,percentage_errors
The Kaggle Master,"Prediction of RNA Binding Sites in a Protein
     CONTEST OVERVIEW:

This contest is going to be a beginner level Kaggle challenge and this is a great opportunity to take a step forward if you are a beginner in Machine Learning and Data Science. All you need basic python knowledge and a zeal to learn. If you devote the required amount of time and efforts, by the end of this contest youll realise that youve grown a lot.

PROBLEM OVERVIEW:

Proteins are the molecular workforces in living organisms. They perform a broad range of essential functions. They catalyze metabolic reactions, replicate DNA, respond to stimuli, provide movement, and much more. 
So in this challenge, we will try to look upon a specific type of protein named RBPs                 (RNA Binding Proteins) with a specific function i.e to bind to a specific target site on RNA. RBPs bind to specific target sites, how they identify these sites is still under research. Understanding how these proteins bind to a specific RNA is of great importance as it can provide us with another way to identify target proteins with similar functions. So the identification of RBPs and their binding sites is a major challenge in the field of molecular recognition. Determining the RNA interacting residues in a protein from its structure is quite easy but it is time-consuming and costly. 

So we want to develop a machine learning model to predict RNA binding sites of a protein from its amino acid sequence. 


Topics you may want to read on your own:
Basic knowledge about proteins.
Scoring Matrices


EVALUATION:

Participants will be evaluated based on the Area Under Curve (AUC) of ROC for their Kaggle Submissions.
Plagiarism will not be tolerated and will result in disqualification.


NOTE:

BEWARE OF OVERFITTING!!


REFERENCES:
Hint related to how to build an appropriate input matrix - To capture the effects caused by evolution in the amino acid sequence PSSM matrix can be used. (https://www.cs.rice.edu/~ogilvie/comp571/2018/09/11/pssm.html)

A good Applied  Machine Learning course for Beginners: https://www.udacity.com/course/intro-to-machine-learning--ud120

Wondering how to apply classification techniques to your input data you can refer to https://towardsdatascience.com/building-classification-models-with-sklearn-6a8fd107f0c1 . ",tabular data,inclass,Prediction of RNA Binding Sites in a Protein,auc,the-kaggle-master,auc
Basic Regression Competition,"This ""code"" competition invites you to train a system that will mimic an unknown function with 8 variables:
f(a,b,c,d,e,f,g,h) = ?
Each variable is an integer in the range (1, 20) inclusive. The output is an integer which will always be within the range (1, 2000) (but the actual range of values is likely smaller). You get 500 samples to train on and need to predict 500 new values.
Extra Details
I wrote a simple function that takes 8 inputs and generates an output value by adding 8 independent mini-functions and then rounding the result to an integer.
So in reality it looks something like this:
f(a,b,c,d,e,f,g,h) = round(fn2(a,f) + fn2(b) + fn3(c,d,e) + ...)
Not every term is weighted equivalently. Some terms are simple and heavily weighted. Some terms are complex and less heavily weighted. This should allow many systems to find a relatively good answer relatively easily (but it might take some extra work to get a score of 0).
Operators
The operators used in the functions are operations that would be considered basic. Addition, subtraction, multiplication, division plus some basic built-in operators likesqrt, sin, cos, etc.",tabular data,inclass,Find the best fit for a non-linear regression problem with 8 input variables and one output variable,mae,basic-regression-competition,mae
Brain Cancer Classification,"Context
This dataset is about brain cancer gene expression. Brain cancers are malignant brain tumors which is a growth of abnormal cells that have formed in the brain [ABTA].
Gene expression profiling is a laboratory method that identifies all of the genes in a cell or tissue that are making messenger RNA. Messenger RNA molecules carry the genetic information that is needed to make proteins from the DNA in the nucleus of the cell to the cytoplasm where the proteins are made. A gene expression profile may be used to find and diagnose a disease or condition or to see how well the body responds to treatment [NCI].
The dataset contains 4 types of brain tumors: ependymoma, glioblastoma, medulloblastoma, and pilocytic astrocytoma.
Content
Original data came from CuMiDa: An Extensively Curated Microarray Database via Kaggle Datasets.
The downloaded dataset contains 130 instances (patients) including 46 are ependymoma, 34 are glioblastoma, 22 are medulloblastoma, 15 are pilocytic astrocytoma, and the remaining 13 are normal.
Training Data and Test Data
The training data is randomly chosen 91 (70%) instances.
The test data is the remaining 39 (30%) instances.
License
ODbL 1.0 in Kaggle Datasets.
Notice
This is a private competition in 1056Lab, a data mining laboratory at Chubu University.
References
Bruno Grisci: Brain cancer gene expression - CuMiDa - Kaggle Datasets
CuMiDa: An Extensively Curated Microarray Database - SBCB
Feltes, B.C.; Chandelier, E.B.; Grisci, B.I.; Dorn, M. CuMiDa: An Extensively Curated Microarray Database for Benchmarking and Testing of Machine Learning Approaches in Cancer Research. Journal of Computational Biology, Ahead of Print, 2019.
Brain Tumor Education - American Brain Tumor Association
NCI Dictionary of Cancer Terms - National Cancer Institute",tabular data,inclass,The 19th 1056Lab Data Analytics Competition,meanfscore,brain-cancer-classification,f_score_multiclass
Machine Learning Lab - CAS Data Science FS 20,"Diese Competition ist fr den Corona bedingten Online-Teil des Machine Learning Labs im Rahmen des CAS Data Science der FHNW im Frhlingssemester 2020.
Aufgabe
Wir haben hier einen Datensatz bereitgestellt, welcher Informationen von 22570 Liegenschaften in der Schweiz beinhaltet.
Ziel ist es fr die 2257 Liegenschaften im Testset den Preis zu schtzen.
Mgliches Vorgehen
Daten einlesen
Daten in Trainings- und Development-Set teilen
Daten subern
Daten analysieren
Korrelationsmatrix
Koordinaten
Outliers identifizieren
Einfaches Modell trainieren um Preis zu schtzen (z.B. Lineare Regression)
Modell auf Development-Set evaluieren
Preise auf Test-Set schtzen und Submission hinaufladen
Weitere Modelle trainieren und evaluieren
Gradient Boosting
Verschiedene Modelle fr die verschiedenen Objekttypen trainieren
Ensemble mehrere Modelle trainieren
Feature Engineering
Features transformieren
Features kombinieren
neue Features suchen (z.B. Steuerfuss der Gemeinde, Geo Daten)
Hyperparameter optimieren",tabular data,inclass,FHNW Edition,mape,machine-learning-lab-cas-data-science-fs-20,percentage_errors
Challenge GH,"Welcome
Willkommen bei Ihrer persnlichen Data Analytics Challenge.
Viel Erfolg und vor allen Dingen viel Spa!
Aufgabenstellung
Sie schlpfen in die Rolle eines Data Analysten der MedBank. Einer auf den Gesundheitsmarkt spezialisierten Genossenschaftsbank.
Sie haben soeben Ihr neues Projekt und den zugehrigen, Ihnen unbekannten, Datensatz erhalten.
Ihre Aufgabe ist es, ein mglichst gutes Prognosemodell fr eine Marketingkampagne fr eine neues Produkt ""MedTrust"" zu erstellen und die Abschlusswahrscheinlichkeiten vorauszusagen.
Analysieren Sie die hierfr den Trainingsdatensatz.
Fhren Sie zunchst eine explorative Analysen durch: Fr welche der erklrenden Variablen sind Zusammenhnge zur Zielvariablen zu beobachten? Welche sind fachlich sinnvoll?
Treffen Sie eine Auswahl von erklrenden Variablen und erstellen Sie anschlieend verschiedene Vorhersagemodelle (logistische Regression, Entscheidungsbume sowie Random Forests).
Erstellen Sie je Verfahren verschiedene Modelle mit unterschiedlichen Parametern. Gehen Sie hierbei mglichst systematisch vor, so dass Sie spter die Performance des Modells (=AUC) in Abhngigkeit von den Parametern beschreiben knnen.
Wenden Sie dieses Modell an, um eine Prognose den Testdaten vorzunehmen.",tabular data,inclass,Challenge GH,auc,challenge-gh,auc
Tracy Regression,"This is the home page of the competition. The first phase will be to use multiple regression to predict diamond price. This phase will use the lm() function.
Phase 2
The next phase of the competition will be to redo the approach from phase 1 in Python. You can use the same algorithm or tweak it, based on what you learned.
Phase 3
This phase will allow use of other functions other than lm(). This will give you a chance to explore boosting, neural networks, etc.
Phase 4
Phase 4 will allow use of stacking and ensembling techniques to improve your model.
Dataset Overview
Here is a decent analysis of the diamonds dataset.
https://www.kaggle.com/stansilas/shine-bright-like-a-diamond/notebook
We thank developers of the ggplot2 library for providing this dataset.",tabular data,inclass,Competition to explore multiple regression for Tracy High School,rmsle,tracy-regression,percentage_errors
Data Science - Master,"Cette page a pour but de dcrire le jeu de donnes ""Booking.com"".
Vous devez tre capable de rpondre aux questions suivantes : Quel fichier dois-je choisir ? Quel est le format des features ? Quelle est la variable  prdire ?
Files descriptions
Attention : le sparateur est la virgule !
train_booking.csv - Jeu de donnes dapprentissage
test_booking.csv - Jeu de donnes de test
submission_booking.csv - Un exemple du fichier  soumettre sur Kaggle
Data fields
Format de lecture : colonne - description colonne - format de colonne
train_booking.csv
id - Identifiant de lannonce - string
name - Nom de lannonce - string
description - Description de lannonce- string
neighborhood_overview - Description du quartier- string
notes - Remarques sur le logement- string
transit - Moyens de locomotion proche du logement - string
access - Accs au logement- string
intercation - Interaction avec lhte- string
house_rules - Rgles dans le logement- string
host_id - Identifiant de lhte- integer
host_since - Date dinscription de lhte - Date
host_id - Identifiant de lhte- integer
host_seniority - Anciennet en jour de lhte (depuis son inscription) - integer
hostresponsetime - Temps de rponse de lhte  une question - String
hostresponserate - Taux de rponse de lhte aux questions - String
hostlistingscount - Nombre dannonces de lhte sur ""Booking.com"" - integer
host_verifications - Vrifications sur lhte - String
neighbourhood_cleansed - Quartier le plus proche - String
city - Ville de lannonce - String
zipcode - Code postal de lannonce - Integer
country - Pays de lannonce - String
latitude_booking /strong> - Latitude de lannonce - Decimal
longitude_booking - Longitude de lannonce - Decimal
property_type - Type de logement - String
room_type - Type de chambre - String
accommodates - Nombre de pices - Integer
bathrooms - Nombre de salles de bains - Integer
bedrooms - Nombre de chambres - Integer
beds - Nombre de lits - Integer
bed_type - Type de lit - String
amenities - Equipements - String
price - Prix dune nuit dans le logement - Integer
security_deposit - Dpt de garantie - String
cleaning_fee - Frais de nettoyage - String
guests_included - Nombre dinvits spciaux autoriss - Integer
extra_people - Supplment () si prsence dinvits spciaux - Integer
minimum_nights - Nombre minimal de nuits autorises - Integer
maximum_nights - Nombre maximal de nuits autorises - Integer
calendar_updated - Calendrier de lannonce mis  jour - String
availability_30 - Nombre de nuits possibles pour les 30 prochains jours - Integer
availability_60 - Nombre de nuits possibles pour les 60 prochains jours - Integer
availability_90 - Nombre de nuits possibles pour les 90 prochains jours - Integer
availability_365 - Nombre de nuits possibles pour les 365 prochains jours - Integer
first_review - Date du premier commentaire - Date
last_review - Date du dernier commentaire - Date
reviewscoresrating -Nombre de commentaires ayant donn la note maximale sur le total de - commentaires ayant donn une note - Integer
reviewscoresaccuracy - Score moyen sur la description du logement - Integer
reviewscorescleanliness - Score moyen sur la propret du logement - Integer
reviewscorescheckin - Score moyen sur le check-in du logement - Integer
reviewscorescommunication - Score moyen sur la communication du logement - Integer
reviewscoreslocation - Score moyen sur la localisation du logement - Integer
reviewscoresvalue - Score moyen du logement - Integer
cancellation_policy - Type dannulation pour la rservation - String
reviewspermonth - Nombre de commentaires par mois - Decimal
geolocation - Golocalisation du logement (latitude, longitude) - String
geopoint_announce - Golocalisation du logement - GeoPoint
department - Dpartement du logement - String
submission_booking.csv
Id - Identifiant de lannonce - string
Predicted - Estimation du prix dune nuit dans le logement - Double
Extranal Data
Stations RATP avec le fichier ""externaldataratp.csv""
latituderatp - Latitude - Double longituderatp - Longituderatp - String idrefzdl - Identifiant station - Integer garesid - Identifiant de la gare - Integer
nomlong - Nom de la station v1 - String
nummod - Numro modifi - Integer fer - Flag fer- Integer train - Flag train- Integer rer - Flag rer - Integer metro - Flag mtro - Integer tramway - Flag tramway - Integer navette - Flag navette - Integer val - Flag val- Integer rescom - Nom de la station v2 - String
geopoint - GoPoint de la station - GeoPoint",tabular data,inclass,Modélisation Data Science des annonces Booking.com,rmse,data-science-master,mse
CSC: HW4 spring19,"          ,           .
:         0/1.
   >2    Nave,    5 --   Advanced   MAP@1.
    >2   ,   leaderboard   MAP@1: ,    CS center   kernels   ,    .",tabular data,inclass,"И снова деревья, ансамбли, ...",auc,csc:-hw4-spring19,auc
Flatiron School,"You have just received a hologram message from a droid:
Future Data Science Jedi Knight! Years ago, you served with my father in the Clone Wars.  Now he begs you to help him in his struggle against the Empire. The Rebellion is under siege from a new division of the Empire  propaganda video games!  We need you to predict the sales of a collection of their video games. Well use this information to bring peace to the Galaxy.

This is our most desperate hour.  Help me, Future Data Science Jedi Knight. Youre my only hope.  [looks to the side quickly, then crouches to end the message]
Best of luck to everyone and dont forget your Jedi Knight Training:
Be Scrappy
Radiate Positivity
Pursue Mastery
Work Together
Make No Little Plans",tabular data,inclass,Use your regression skills to save the Galaxy,mse,flatiron-school,mse
Technidus machine learning competition 2,"This is the second competition to test your understanding of machine learning concepts learnt. For this competition, you will be applying your knowledge about algorithms and modelling to predict a customers purchase behaviour. Use the customer data provided to create a classification model that predicts a customers likelihood to buy a bike.
Ensure you apply what youve learned in the resources shared.",tabular data,inclass,Making predictions on customer purchase behaviour,categorizationaccuracy,technidus-machine-learning-competition-2,categorization_accuracy
Data Champions Android App Malware Prediction,"Welcome Data Champions
Youve seen a small piece of the Android Permissions dataset but now youre getting all of it. This challenge tests what youve learnt in a competitive environment with a public and private leaderboard.
The aim is predict if a given Android application is Malware or not. Youll be given 14889 labeled observations which you can use for training and validation and 6382 unlabeled observations which will be used for testing. Your position on the public leaderboard will depend on your predictions on 70% of the test data and the remaining 30% of the test dataset will be used to determine the private leaderboard which will only be revealed after the competition has ended.
Where do I start?
The first thing to do is read through the Data documentation and then download the data. Once you have saved and extracted it, you can read it into Python using Pandas and split the training data into Train and Validation datasets. Or you can leave the training data as is and use Cross Validation. Or you can do both, you can experiment and decide.
Once you have the data loaded, you can start exploring and building models! Just follow the Data Champions lessons
Acknowledgements
The data are licensed under a Creative Commons Attribution 4.0 International licence. Changes were made to the data to prevent using the original data to cheat in this competition. We thank the data author, Arvind Mahindru and provide a DOI to the original dataset 10.17632/958wvr38gy.5",tabular data,inclass,"Using permissions for Android applications, predict whether or not the application is Malware or Benign",auc,data-champions-android-app-malware-prediction,auc
DSI-US-8 Project 2 Regression Challenge,"Welcome the Kaggle challenge for Project 2! As part of a successful submission for Project 2, we will expect you to make at least one andhopefully,multiple! submissions towards this regression challenge.
In this challenge, you will use the well known Ames housing data to create a regression model that predicts the price of houses in Ames, IA. You should feel free to use any and all features that are present in this dataset.
Goal
It is your job to predict the sales price for each house. For each Id in the test set, you must predict the value of the SalePrice variable.
Evaluation
Kaggle leaderboard standings will be determined by root mean squared error RMSE.

$$\text{RMSE}=\sqrt{\sum\frac{(\hat{y}i - yi)^2}{n}}$$
Submission File Format
The file should contain a header and have the following format:.

Id,SalePrice",tabular data,inclass,Predict the price of homes at sale for the Aimes Iowa Housing dataset,rmse,dsi-us-8-project-2-regression-challenge,mse
InClass Competition at Tokyo Metropolitan Univ.,,tabular data,inclass,Predicting the price for an Airbnb host.,rmsle,inclass-competition-at-tokyo-metropolitan-univ.,percentage_errors
Exam for Students20200129,"AI Academy
SubmissionKernel (Python or R)
csvKernelnotebooksubmit
KernelPublic
Kernel
Rules


EvaluationRMSLE
",tabular data,inclass,This is Exam for students in the Academy. Have fun with ML!,rmsle,exam-for-students20200129,percentage_errors
Penyisihan Datavidia 2019,"Data untuk kompetisi ini diberikan oleh salah satu ticketing company terbesar di Indonesia
Pada kompetisi ini, peserta diminta untuk memprediksi apakah akan terjadi cross selling pada transaksi flight seorang pelanggan. Cross selling didefinisikan sebagai:
Pelanggan membuat booking hotel bersamaan dengan membeli tiket pesawat
Selamat mengerjakan dan selamat berkompetisi!",tabular data,inclass,"Babak penyisihan untuk Datavidia, hosted by Arkavidia 6.0",meanfscore,penyisihan-datavidia-2019,f_score_multiclass
Predictive Equipment Failures,"Low Margins, High Importance
Background:
80% of producing oil wells in the United States are classified as stripper wells. Stripper wells produce low volumes at the well level, but at an aggregate level these wells are responsible for a significant percentage of domestic oil production.
Stripper wells are attractive to a company due to their low operational costs and low capital intensity - ultimately providing a source of steady cash flow to fund operations that require more funds to get off the ground.
At ConocoPhillips, our West Texas Conventional operations serve as a source of organic cash flow to fund more expensive projects in the Delaware Basin and other unconventional plays across the United States. As a company, it is vital that this steady, low cost form of cash has a constant presence.
As with all mechanical equipment, things break and when things break money is lost in the form of repairs and lost oil production. When costs go up cash goes down, but how can we predict when equipment will fail and use this information to drive down our costs?
The Challenge:
A data set has been provided that has documented failure events that occurred on surface equipment and down-hole equipment. For each failure event, data has been collected from over 107 sensors that collect a variety of physical information both on the surface and below the ground.
Using this data, can we predict failures that occur both on the surface and below the ground? Using this information, how can we minimize costs associated with failures?
The goal of this challenge will be to predict surface and down-hole failures using the data set provided. This information can be used to send crews out to a well location to fix equipment on the surface or send a workover rig to the well to pull down-hole equipment and address the failure.
In addition to uploading a solution file (described in ""Evaluation""), teams will be asked to provide a ""kernel"" via a markdown file. The kernel provides us with your code and output in addition to answers for the prompts in the ""Kernels Requirements"" section. These prompts in the ""Kernels Requirements"" section will determine your overall placement in the competition.",tabular data,inclass,Predict downhole equipment failures using sensor data!,meanfscore,predictive-equipment-failures,f_score_multiclass
CSM/SEM6420 workshop,"Description. Your task for this exercise is to develop machine learning models of your own choice to predict the likelihood of blood-brain barrier penetration for a chemical compound, based on the given chemical structural information.
Evaluation. The Area Under the Receiver Operator Characteristic (AUC) will be used for evaluation (see http://scikit-learn.org/stable/modules/generated/sklearn.metrics.roc_auc_score.html and https://en.wikipedia.org/wiki/Receiver_operating_characteristic for details about this metric. 
Submission. The submitted csv file should consist of two columns with the header of: ""TestId"" and ""PredictedScore"" (i.e. probability output or score for a chemical being BBB permeable). 
Leaderboard and final evaluation. The predictions on about 50% of the test data points are used to score the submission according to the AUC and maintain a public leaderboard. The predictions on the remaining test data points will be used, after the submission deadline, for the final evaluation and you receive marks according to the resulting area under the ROC curve (AUC) for the remaining test set. This prevents a high score in the final evaluation from being obtained through overfitting the public test data, but it means that the public leaderboard will not necessarily be indicative of final performance.
Acknowledgements
The original dataset can be found in https://pubs.acs.org/doi/suppl/10.1021/ci300124c
The publication associated with this data is:
Martins IF, Teixeira AL, Pinheiro L, Falcao AO. (2012). A Bayesian approach to in silico blood-brain barrier penetration modeling. Journal of Chemical Information and Modeling, 52(6),1686-97.
We thank the authors with University of Lisbon for providing the original dataset.",tabular data,inclass,Compound classification using chemical structural information,auc,csm/sem6420-workshop,auc
Eval Lab 2 F464,"Your task is to predict type of toy from given features. Accuracy will be used to evaluate your predictions. When you submit your predictions only 50% of them will be evaluated and you can see your score on the Public Leaderboard. At the end of the contest the remaining 50% will be evaluated and final standings will be displayed on Private Leaderboard. Simple average of Public and Private Leaderboards will be considered as final score.
This is competition is evaluative.",tabular data,inclass,The second evaluative lab of BITS F464 Fall 2019,categorizationaccuracy,eval-lab-2-f464,categorization_accuracy
finec-1941-hw6,"  6.    .
   10   , 2  .   -    (K-means, DBSCAN, etc)      .
:
numpy
pandas
sklearn.model_selection
sklearn.metrics
sklearn.multiclass
sklearn.preprocessing
scipy
:   
       :  -     https://www.kaggle.com/rounakbanik/the-movies-dataset , . https://www.kaggle.com/c/finec-1941-hw6/data
  ,   private kernel   ;  (@gorodec)   ,  ",tabular data,inclass,Clustering,macrofscore,finec-1941-hw6,f_score_multiclass
Data Series Summarization Project (v3),"Our first task is to provide a summarization for all the provided 50k time series of both synthetic dataset and seismic dataset.
Informally, summarization means to use less memory to represent the original dataset. This means usually summarization could bring space efficiency. For example, the size of our original dataset is 4 byte / float * 256 float numbers * 50000 series = 51.2 MB, with each series occupying 1024 bytes.
Our summarization should be using less memory. In this project, we will need to summarize each series (of 1024 bytes) using
 32 bytes
 64 bytes
 128 bytes
We will evaluate the effectiveness of our summarizations for these 3 different cases.
Notebook
Submissions have to be in Python
Go to Notebooks, then Your Work, then Create New Notebook (or duplicate the template notebook)
The submitted notebook need to have 3 functions for summarization and 3 functions for reconstruction with the exacts following names:
sum32
sum64
sum128
rec32
rec64
rec128
Each summarization function take as an input argument the path of the file containing 50000 series (a series contains 256 data points of 4 bytes each). Each summarization function should calculate the summarization of the 50000 series, write the results into a file, then return the path of the summarization file named by concatenating to its name _sum32, _sum64, or _sum128 respectively.
For instance, if the filename is syntheticsize50klen256znorm.bin, then the sum32 function should output filename path as syntheticsize50klen256znorm.bin_sum32.
Each reconstruction function take as an input argument the path of the file containing 50000 summaries (of respectively 32 bytes, 64 bytes, or 128 bytes). Each reconstruction function should calculate the reconstruction of the 50000 series summaries, write the results into a file, then return the path of the reconstruction file named by concatenating to its name _rec32, _rec64, or _rec128 respectively.
A solution template can be found here https://www.kaggle.com/abdumaa/submission-template-time-series
Submission
The submission file is a CV file containing a header id, expected followed by the reconstructed time series based on 32-bytes, 64-bytes, and 128-bytes summaries as the following:
Indexes of the time series (back to back ), based on 32-bytes summaries, span from id=0 to id =12799999.
Indexes of the time series (back to back ), based on 64-bytes summaries, span from id=12800000 to id =25599999.
Indexes of the time series (back to back ), based on 128-bytes summaries, span from id=25600000 to id =38399999.
Here the official documentation to submit a solution from a notebook
https://www.kaggle.com/dansbecker/submitting-from-a-kernel
I would rather recommend this youtube video of how to participate and submit a solution in a kaggle competition from a notebook (kernel)https://www.youtube.com/watch?v=GJBOMWpLpTQ
To follow up on the youtube video, if you cannot find the button ""Commit"", so you need to click on ""Save Version"", then choose ""Save & Run All (commit)""
Do not forget to share your submission notebook after you have submitted your solution:
click on the button ""share"",
type ""abdu maa"" into the ""collaborator"" search box
click ""save""
Since the NAMES/TITLES of submissions and shared notebooks for both the summarization and similarity Kaggle competitions are NOT INFORMATIVE, please make sure to include you submission and notebook title in the online Google sheet https://docs.google.com/spreadsheets/d/1mtVwiG9aelYIeS44Aoz50396p2WCywCBE2aVTOvbut0/edit#gid=1976504894",tabular data,inclass,"In this project, we will explore one important technique of time series analysis, named summarization.",rmse,data-series-summarization-project-(v3),mse
UCI-HAR,"===================================================================================================
Human Activity Recognition Using Smartphones Dataset
Version 1.0
Jorge L. Reyes-Ortiz(1,2), Davide Anguita(1), Alessandro Ghio(1), Luca Oneto(1) and Xavier Parra(2)
1 - Smartlab - Non-Linear Complex Systems Laboratory
DITEN - Universit degli Studi di Genova, Genoa (I-16145), Italy.
2 - CETpD - Technical Research Centre for Dependency Care and Autonomous Living
Universitat Politcnica de Catalunya (BarcelonaTech). Vilanova i la Geltr (08800), Spain
activityrecognition @ smartlab.ws
The experiments have been carried out with a group of 30 volunteers within an age bracket of 19-48 years. Each person performed six activities (WALKING, WALKINGUPSTAIRS, WALKINGDOWNSTAIRS, SITTING, STANDING, LAYING) wearing a smartphone (Samsung Galaxy S II) on the waist. Using its embedded accelerometer and gyroscope, we captured 3-axial linear acceleration and 3-axial angular velocity at a constant rate of 50Hz. The experiments have been video-recorded to label the data manually. The obtained dataset has been randomly partitioned into two sets, where 70% of the volunteers was selected for generating the training data and 30% the test data.
The sensor signals (accelerometer and gyroscope) were pre-processed by applying noise filters and then sampled in fixed-width sliding windows of 2.56 sec and 50% overlap (128 readings/window). The sensor acceleration signal, which has gravitational and body motion components, was separated using a Butterworth low-pass filter into body acceleration and gravity. The gravitational force is assumed to have only low frequency components, therefore a filter with 0.3 Hz cutoff frequency was used. From each window, a vector of features was obtained by calculating variables from the time and frequency domain. See features_info.txt for more details.
For each record it is provided:
Triaxial acceleration from the accelerometer (total acceleration) and the estimated body acceleration.
Triaxial Angular velocity from the gyroscope.
A 561-feature vector with time and frequency domain variables.
Its activity label.
An identifier of the subject who carried out the experiment.
The dataset includes the following files:
README.txt
features_info.txt: Shows information about the variables used on the feature vector.
features.txt: List of all features.
activity_labels.txt: Links the class labels with their activity name.
train/X_train.txt: Training set.
train/y_train.txt: Training labels.
test/X_test.txt: Test set.
test/y_test.txt: Test labels.
The following files are available for the train and test data. Their descriptions are equivalent.
train/subject_train.txt: Each row identifies the subject who performed the activity for each window sample. Its range is from 1 to 30.
train/Inertial Signals/totalaccxtrain.txt: The acceleration signal from the smartphone accelerometer X axis in standard gravity units g. Every row shows a 128 element vector. The same description applies for the totalaccxtrain.txt and totalaccz_train.txt files for the Y and Z axis.
train/Inertial Signals/bodyaccx_train.txt: The body acceleration signal obtained by subtracting the gravity from the total acceleration.
train/Inertial Signals/bodygyrox_train.txt: The angular velocity vector measured by the gyroscope for each window sample. The units are radians/second.
Notes:
Features are normalized and bounded within [-1,1].
Each feature vector is a row on the text file.
The units used for the accelerations (total and body) are gs (gravity of earth -> 9.80665 m/seg2).
The gyroscope units are rad/seg.
A video of the experiment including an example of the 6 recorded activities with one of the participants can be seen in the following link: http://www.youtube.com/watch?v=XOEN9W05_4A
For more information about this dataset please contact: activityrecognition @ smartlab.ws
License:
Use of this dataset in publications must be acknowledged by referencing the following publication [1]
[1] Davide Anguita, Alessandro Ghio, Luca Oneto, Xavier Parra and Jorge L. Reyes-Ortiz. A Public Domain Dataset for Human Activity Recognition Using Smartphones. 21th European Symposium on Artificial Neural Networks, Computational Intelligence and Machine Learning, ESANN 2013. Bruges, Belgium 24-26 April 2013.
This dataset is distributed AS-IS and no responsibility implied or explicit can be addressed to the authors or their institutions for its use or misuse. Any commercial use is prohibited.
Other Related Publications:
[2] Davide Anguita, Alessandro Ghio, Luca Oneto, Xavier Parra, Jorge L. Reyes-Ortiz. Energy Efficient Smartphone-Based Activity Recognition using Fixed-Point Arithmetic. Journal of Universal Computer Science. Special Issue in Ambient Assisted Living: Home Care. Volume 19, Issue 9. May 2013
[3] Davide Anguita, Alessandro Ghio, Luca Oneto, Xavier Parra and Jorge L. Reyes-Ortiz. Human Activity Recognition on Smartphones using a Multiclass Hardware-Friendly Support Vector Machine. 4th International Workshop of Ambient Assited Living, IWAAL 2012, Vitoria-Gasteiz, Spain, December 3-5, 2012. Proceedings. Lecture Notes in Computer Science 2012, pp 216-223.
[4] Jorge Luis Reyes-Ortiz, Alessandro Ghio, Xavier Parra-Llanas, Davide Anguita, Joan Cabestany, Andreu Catal. Human Activity and Motion Disorder Recognition: Towards Smarter Interactive Cognitive Environments. 21th European Symposium on Artificial Neural Networks, Computational Intelligence and Machine Learning, ESANN 2013. Bruges, Belgium 24-26 April 2013.
==================================================================================================
Jorge L. Reyes-Ortiz, Alessandro Ghio, Luca Oneto, Davide Anguita and Xavier Parra. November 2013.",tabular data,inclass,UCI-HAR,categorizationaccuracy,uci-har,categorization_accuracy
UMUC DATA 650 Summer 2019 Competition,"1. Title: Car Evaluation Database
2. Sources:
(a) Creator: Marko Bohanec
(b) Donors: Marko Bohanec (marko.bohanec@ijs.si); Blaz Zupan (blaz.zupan@ijs.si)
(c) Date: June, 1997
3. Past Usage:
The hierarchical decision model, from which this dataset is derived, was first presented in M. Bohanec and V. Rajkovic: Knowledge acquisition and explanation for multi-attribute decision making in 8th Intl Workshop on Expert Systems and their Applications, Avignon, France. pages 59-78, 1988.
Within machine-learning, this dataset was used for the evaluation of HINT (Hierarchy INduction Tool), which was proved to be able to completely reconstruct the original hierarchical model. This, together with a comparison with C4.5, is presented in B. Zupan, M. Bohanec, I. Bratko, J. Demsar: Machine learning by function decomposition. ICML-97, Nashville, TN. 1997 (to appear)
4. Relevant Information Paragraph:
Car Evaluation Database was derived from a simple hierarchical decision model originally developed for the demonstration of DEX (M. Bohanec, V. Rajkovic: Expert system for decision making. Sistemica 1(1), pp. 145-157, 1990.). The model evaluates cars according to the following concept structure:
CAR car acceptability
. PRICE overall price
. . buying buying price
. . maint price of the maintenance
. TECH technical characteristics
. . COMFORT comfort
. . . doors number of doors
. . . persons capacity in terms of persons to carry
. . . lug_boot the size of luggage boot
. . safety estimated safety of the car
Input attributes are printed in lowercase. Besides the target concept (CAR), the model includes three intermediate concepts: PRICE, TECH, COMFORT. Every concept is in the original model related to its lower level descendants by a set of examples (for these examples sets see http://www-ai.ijs.si/BlazZupan/car.html).
The Car Evaluation Database contains examples with the structural information removed, i.e., directly relates CAR to the six input attributes: buying, maint, doors, persons, lug_boot, safety.
Because of known underlying concept structure, this database may be particularly useful for testing constructive induction and structure discovery methods.
Acknowledgement
@misc{Lichman:2013 ,
author = ""M. Lichman"",
year = ""2013"",
title = ""{UCI} Machine Learning Repository"",
url = ""http://archive.ics.uci.edu/ml"",
institution = ""University of California, Irvine, School of Information and Computer Sciences"" }",tabular data,inclass,Evaluate the Cars!,categorizationaccuracy,umuc-data-650-summer-2019-competition,categorization_accuracy
IESB - 2019,"Os dados so do IGM, e iremos prever se a nota de matemtica de um municpio na prova de matemtica do ENEM est acima ou abaixo da mediana Brasil.
As notas sero calculadas de acordo com os seguintes critrios:
Nota mais alta na competio
Melhor Kernel de anlise exploratria na competio
Melhor Kernel de anlise de feature importances
Cronograma:
21/05 - Aula 06 - Elaborao de Modelo do Kaggle
28/05 - Aula 07 - Apresentao dos modelos",tabular data,inclass,Machine Learning Hands-on Course,rmse,iesb-2019,mse
MLDM Classification Competition,"Wine (from Latin vinum) is an alcoholic beverage made from grapes, generally Vitis vinifera, fermented without the addition of sugars, acids, enzymes, water, or other nutrients.
Wine has been produced for thousands of years. The earliest known traces of wine are from Georgia (cca. 6000 BC), Iran (cca. 5000 BC), and Sicily (cca. 4000 BC) although there is evidence of a similar alcoholic beverage being consumed earlier in China (cca. 7000 BC). The earliest known winery is the 6,100-year-old Areni-1 winery in Armenia. Wine reached the Balkans by 4500 BC and was consumed and celebrated in ancient Greece, Thrace and Rome. Throughout history, wine has been consumed for its intoxicating effects.
Wine has long played an important role in religion. Red wine was associated with blood by the ancient Egyptians and was used by both the Greek cult of Dionysus and the Romans in their Bacchanalia; Judaism also incorporates it in the Kiddush and Christianity in the Eucharist.
Yeast consumes the sugar in the grapes and converts it to ethanol and carbon dioxide. Different varieties of grapes and strains of yeasts produce different styles of wine. These variations result from the complex interactions between the biochemical development of the grape, the reactions involved in fermentation, the terroir, and the production process. Many countries enact legal appellations intended to define styles and qualities of wine. These typically restrict the geographical origin and permitted varieties of grapes, as well as other aspects of wine production. Wines not made from grapes include rice wine and fruit wines such as plum, cherry, pomegranate and elderberry.
This work addresses the following issues concerning the quality of wine with respect to various chemical contents of and acids. The first issue concerns the correlation between different acids and quality of wine. In this work, we investigate what could be the best associate parameter on which best quality wine depends?
Overview of the Study
Our field study concerns quality of wine , produced all over the world. Wine is from one of the alcohol family and it is considered as a part of rich culture as well. There are various health benifits of wine (http://www.wideopeneats.com/10-health-benefits-get-drinking-daily-glass-wine/). There are a large number of occupations and professions that are part of the wine industry, ranging from the individuals who grow the grapes, prepare the wine, bottle it, sell it, assess it, market it and finally make recommendations to clients and serve the wine. In this study, we figure out important correlated chemical components of wine. Important acids which are associated with quality of wine. We will find out the right components which are needed in right ratio to make quality wine.",tabular data,inclass,"The dataset is related to red and white variants of the Portuguese ""Vinho Verde"" wine. ",meanfscore,mldm-classification-competition,f_score_multiclass
IES Data Mining(WS 18/19),"Diese freiwillige Semesteraufgabe dient dazu, Ihr erlerntes Wissen zu zeigen. Durch Bearbeiten der Aufgabe knnen Sie sich alleine oder in Gruppen von maximal 3 Personen einen Bonus fr die Prfung am Ende des Semesters erarbeiten. Am Ende der Vorlesungszeit wird ein Termin angesetzt, an dem Sie Ihre Lsung kurz vorstellen, und die beste Lsung mit einem Preis prmiert wird.
In der Semesteraufgabe geht es um die Klassifikation des gegebenen Datensatzes. Der Datensatz besteht dabei aus verschiedenen Datentypen:
kategorisch
kontinuierlich
missing values
u.v.m.
Ihre Aufgabe besteht darin ein Modell zu entwerfen, das die Daten den entsprechenden Klassen zuordnet. Sie knnen den kompletten Datensatz zum Training und Testen verwenden. Bei der Vorstellung Ihres Algorithmus werden wir Ihre Lsung mit neuen, unbekannten Daten testen.
Die Ergebnisse der einzelnen Gruppen werden mit Hilfe von Precission, Recall und dem F1-Measure miteinander verglichen.",tabular data,inclass,Student Competition,meanfscore,ies-data-mining(ws-18/19),f_score_multiclass
HEROZ Internal Competition,"Airbnb(price)
RMSE
AirbnbKaggle1CPU(1Core)+1GPU
Private Leaderboard
submit20",tabular data,inclass,HEROZ社内コンペティション,rmse,heroz-internal-competition,mse
HEROZ Internal Competition Extra2,"Airbnb(price)
RMSE
AirbnbKaggle1CPU(1Core)+1GPU
Private Leaderboard
submit20",tabular data,inclass,HEROZ社内コンペ延長線(fixed),rmsle,heroz-internal-competition-extra2,percentage_errors
Homework for Students,"Feature EngineeringBuild Models
SubmissionNotebooks
csvKernelnotebooksubmit
KernelPublic
Kernel
Forum
Rules",tabular data,inclass,Happy modeling!,auc,homework-for-students,auc
"IEEE PES BDC DataThon , Year-2020","The Challenge
The competition is simple: use machine learning to create a model that predicts Global Horizontal Solar Irradiance (GHI) from a set of features. Please see the ""Rules"" tab before participating.
Overview of How Kaggles Competitions Work
Join the Competition
Read about the challenge description, accept the Competition Rules and gain access to the competition dataset.
Get to Work
Download the data, build models on it locally or on Kaggle Kernels (our no-setup, customizable Jupyter Notebooks environment with free GPUs) and generate a prediction file.
Make a Submission
Upload your prediction as a submission on Kaggle and receive an accuracy score.
Check the Leaderboard
See how your model ranks against other competitors on the leaderboard.
Improve Your Score
Check out the Kaggle discussion forum to find lots of tutorials and insights from other competitors.
What Data Will I Use in This Competition?
In this competition, youll gain access to two similar datasets -One dataset is titled train.csv and the other is titled test.csv.
Using the patterns you find in the train.csv data, predict the target variable -Global Horizontal Solar Irradiance on data found in test.csv .
Check out the Data tab to explore the datasets even further. Once you feel youve created a competitive model, submit it to Kaggle to see where your model stands on our leaderboard against other competitors.
How to Submit your Prediction to Kaggle
Once youre ready to make a submission and get on the leaderboard:
Click on the Submit Predictions button
Upload a CSV file in the submission file format as highlighted in the Evaluation Page. Youre able to submit maximum 3 submissions a day.
Acknowledgement
Primary source for the Dataset-Bangladesh - Solar Radiation Measurement Data
The campaign that has generated this data was commissioned by The World Bank with funding from the Energy Sector Management Assistance Program (ESMAP). The data is made freely available under The World Banks open data policy.
Site Name : BDFE2 (Feni)
Equipment: Helioscale omega station (Tier 1)
Host Institution: Char Darbesh Adarsha Gram Government Primary School
Elevation (m): 5
Latitude (positive North decimal degrees):22.80029
Longitude (positive East decimal degrees):91.35819",tabular data,inclass,What is the value of solar radiation?,rmse,"ieee-pes-bdc-datathon-,-year-2020",mse
Schnell-Mal-Klassifizieren,"Sehr geehrte Studierende,
diese Kaggle-Competition ist nur fr Sie! Lernen Sie hier, mit Hilfe Ihres Klassifikators (z.B. einem k-Nearest-Neighbour-Klassifikators) eine Vorhersage zu treffen.
Unter dem Reiter ""Data"" erhalten Sie die Trainingsdaten in train.csv, um das Target-Label y fr die Daten in test.csv vorherzusagen.
Zum Format der Abgabe gibt es eine Datei SampleSubmission.csv, welche das bentigte Format erlutert. Kurz gesagt ist eine .csv-Datei verlangt der Form
Id,y
0,1
1,0
2,0
3,0
4,1
...
8099,1
Um die Daten zu holen, knnen Sie das Kaggle API benutzen. Dazu mssen Sie das CLI installieren und Ihre Login-Credential in Ihrem Profil holen und unter ~/.kaggle/kaggle.json abspeichern. Danach:
kaggle competitions download -c male-daan-schnell-mal-klassifizieren
Laden Sie Ihre .csv-Datei als Kaggle-Submission hoch!
und
kaggle competitions submit -c male-daan-schnell-mal-klassifizieren -f clever_submission.csv -m ""mein zweiter Versuch heute. Erwartete Genauigkeit 95.3%""
Natrlich geht dies alles auch per Maus und Klick-Klick.
Achtung! Sie haben pro Tag nur die Berechtigung zwei Submissions hochzuladen. Alles andere wird nicht gezhlt. Also berlegen Sie sich gut, welche Prediction Sie hochladen",tabular data,inclass,"Des KNNs Traum, des Perzeptrons 2D-Alptraum",categorizationaccuracy,schnell-mal-klassifizieren,categorization_accuracy
ML Hackathon 2019 Q1,"Welcome to the ML Hackathon 2019, conducted by the Developer Students Club at ASE, Coimbatore !
This dataset contains information collected from various patients records, based on various medical observations and diagnostic tests.
Your task is to determine whether a patient suffers from gastric ulcers or not, based on the given information
Disclaimer
This dataset has been largely fabricated and must NOT be used anywhere else for referral/research/citation purposes or otherwise",tabular data,inclass,Can you predict the presence of a gastric ulcer in a patient? ,categorizationaccuracy,ml-hackathon-2019-q1,categorization_accuracy
MLH - Pokemon Challenge,"Bienvenido a la primera competicin de la comunidad.
Podrs predecir el resultado de futuras batallas pokemon?
Para hacerlo tendrs a tu disposicin las caractersticas de todos los pokemon y los resultados de algunas batallas.
Tres archivos estn disponibles. El primero contiene las caractersticas de Pokmon (la primera columna es el id del pokemon). El segundo contiene informacin sobre batallas anteriores. Las dos primeras columnas son los identificadores del Pokmon y la tercera el identificador del ganador.
Importante: El pokemon en la primera columna ataca primero.
El objetivo es desarrollar un modelo de machine learning capaz de predecir el resultado de futuras batallas pokemon. El mejor modelo gana!",tabular data,inclass,¿Serás capaz de predecir el resultado de los combates pokemon?,categorizationaccuracy,mlh-pokemon-challenge,categorization_accuracy
ML in biology,This is the home page of our first competition. Make you first ML models to predict stroke.,tabular data,inclass,Learn how to use ml methods in biology,auc,ml-in-biology,auc
ML Challenge,"Welcome to Week 6 of the Machine Learning Challenge. This week you get the chance to test the skills that you acquired over the past 5 weeks.
This week will be in the form of a competition. You will have to solve a problem based on a real-world dataset and compete with your colleagues for a place on the leaderboard.",tabular data,inclass,Week 6 - Final Project,auc,ml-challenge,auc
Predicting Age Groups,"Age group prediction challenge
Your task is to predict the age group (1 = ""18-29"", 2 = ""30-49"", 3 = ""50-"") of simulated webshop customers. The webshop uses this prediction to improve its recommendation in case a customer does not provides his/her age. Features are weekday of the transaction, the customers region and the aggregated basket contents of three product groups ""wine"", ""beer"", and ""spirits"".
Use the training dataset to train your machine learning model
Use the test set to create your predictions. You need to export your predictions in the form shown in sample_submission.",tabular data,inclass,"In-class competition for ""pk-hska""",categorizationaccuracy,predicting-age-groups,categorization_accuracy
Machine Learning Lab - CAS Data Science HS 20,"Aufgabe
Wir haben hier einen Datensatz bereitgestellt ber Schler und deren Noten im Fach Mathematik in der Sekundarstufe in Portugal.
Der Datensatz ist eher klein und besteht aus ungefhr 400 Datenpunkte mit jeweil 31 Features und 3 Zielvariabeln. Wir haben den Datensatz in traindata.csv und testdata.csv aufgeteilt, wobei 36 Datenpunkte die Test Daten ausmachen.
Ziel ist es zuerst die Daten zu analysieren und allfllige Probleme fest zu stellen, die wir im Verlauf des Tages beheben knnen. Wir werden diesen Zwischenschritt im Plenum besprechen.
Anschliessend soll ein einfaches Lineares Model auf dem Train Datensatz erstellt werden, welches aus den anderen Zielvariablen G1 und G2 die Zielvariable G3 bestimmt. Dies soll den Einstieg in scikit-learn vereinfachen, da wir vorerst nur zwei numerische Features betrachten. Wir werden diesen Zwischenschritt im Plenum besprechen.
Das soeben erstellte Lineare Modell ist sehr gut, hat aber auch keine schwierige Aufgabe, da G1 und G2 stark mit G3 koreliieren. Deshalb geht es jetzt darum, ein ""richtiges"" Lineares Model zu erstellen, was von den tatschlichen Features auf die Zielvariable G3 geht. Achten Sie dabei auf die Feature Selection und das Encoding von Featuren (One-Hot Encoding). berprfen Sie Ihr Modell zuerst mit einem Validation Set und submitten Sie anschliessend eine Prediction auf Kaggle. Wir werden diesen Zwischenschritt im Plenum besprechen.
Konnten Sie eine Prediction submitten? Bravo! Nun haben Sie freie Hand. Sie knnen beispielsweise Data Cleaning anwenden, andere Modelle ausprobieren (und Hyper Parameter optimieren), Feature Engineering betreiben, weitere Daten anreichern,  - seien Sie kreative.",tabular data,inclass,Mask Edition,mae,machine-learning-lab-cas-data-science-hs-20,mae
Market Basket - ID NDSC 2020,"Description
At Shopee, sellers list thousands of products for sale on our platform. A better understanding of users tastes and preferences for products can help Shopee design better promotions and recommendations for our users. To do that, we conduct market basket analysis which allows us to identify the relationship between different combinations of products that users buy.
We are interested in finding association rules between combinations of different products. These association rules can help to uncover regularities in purchasing behaviors of our users.
For example, an association rule between 3 products, {Product A & Product B}  {Product C}, would indicate that a user buying both Product A & Product B would likely buy Product C as well.
Confidence is a measure that is used to indicate such tendencies and can be used to determine the association for varying numbers of products. For the purpose of this question, we will be using confidence to calculate the association for 2 products and 3 products.
Confidence for two products:
Confidence for three products:
Or
Basic Concepts
Confidence is defined as the tendency that given product A is purchased, that product B will also be purchased.
Each orderid represents a distinct transaction that has occurred.
Each itemid represents a unique product that is sold on Shopee.
A transaction can contain 1 or more itemid(s). If 2 or more itemid(s) share the same orderid, they are purchased together in a single transaction.
An itemid can appear many times in different orderid(s), which means that the product was purchased many times in different transactions.
Task
Please calculate the confidence values for all the association rules provided in the rules.csv file.
Tips:
A > B and B > A have different confidence and should be calculated separately
A & B > C and B & A > C are identical association rules and will yield the same confidence
Example
Case 1: A > B
8 orderid have itemid 7917849
(31338643584868, 31364354557783, 31368958440199, 31369772179043, 31371954695064, 31375314731607, 31377601474289, 31379328498817)
6 orderid out of above have both itemid 7917849 and itemid 18642183
(31338643584868, 31368958440199, 31369772179043, 31371954695064, 31375314731607, 31377601474289)
Confidence (7917849 > 18642183)
= 6 / 8
= 0.75
Please submit 750 (times by 1000 and round down to integer)
Case 2: A&B > C
7 orderid have itemid 2363580843 and itemid 2002243261
(31342449702678, 31365563352719, 31366764361012, 31371701813987, 31372163437582, 31373610230585, 31381568386099)
6 orderid out of above have all itemid 2363580843, itemid 2002243261 and itemid 1993068031
(31342449702678, 31365563352719, 31366764361012, 31372163437582, 31373610230585, 31381568386099)
Confidence (2363580843 & 2002243261 > 1993068031)
= 6 / 7
= 0.857143
Please submit 857 (times by 1000 and round down to integer)
Case 3: A > B&C
9 orderid have itemid 1089203645
(31351735245918, 31367488312991, 31372554805324, 31373458010259, 31373724807962, 31374927925523, 31375318612401, 31375354382289, 31384570619582)
7 orderid out of above have all itemid 1089203645, 431391770 and 1216842899
(31351735245918, 31372554805324, 31373458010259, 31373724807962, 31374927925523, 31375318612401, 31375354382289)
Confidence (1089203645 > 431391770 & 1216842899)
= 7 / 9
= 0.777777
Please submit 777 (times by 1000 and round down to integer)",tabular data,inclass,Challenge #2 for Beginner Category,categorizationaccuracy,market-basket-id-ndsc-2020,categorization_accuracy
NCTU BDALAB 2020 Onboard,"This is a dataset on default payments of credit card clients in Taiwan from April 2005 to September 2005.
We aim to use this competition to try out some basic skills in data science and machine learning.
It will be divided into five main stages, which will span through the three weeks of onboarding.
Trial of Basic Machine Learning Models
Improve Model Performance by doing Exploratory Data Analysis
Considering the imbalanced data problem and Data Augmentation
Dimension Reduction and Feature Selection
Ensemble models (Bagging, Boosting, Stacking)",tabular data,inclass,2020 Onboarding for NCTU BDALAB,auc,nctu-bdalab-2020-onboard,auc
"MLClass Dubai by ODS, Lecture 6 HW","This is a competition for participants of the Open Machine Learning Course by ODS in Dubai.
The goal is to predict musical genres of a track given some features extracted from the wave file and some metadata fields. One track can have several genres.
For this task you can extract any features which you want, and you can use any ML models supported by Kaggle Kernels.
You should write your solutions only in kaggle kernels. You can keep it private until the end of competition, but you if you want you can public any notebook before the end.
Cheating = ban from ODS.
If you cannot show (publish) your notebook after the end of the course, which replicates your score, you will be excluded from the leaderboard.
If you want to get additional credits on the course, you may want to write an article using Kaggle Kernels about F1 score function.
Page of the course",tabular data,inclass,"This is a homework competition for participants of MLClass Dubai by ODS, Lecture 5 HW",meanfscore,"mlclass-dubai-by-ods,-lecture-6-hw",f_score_multiclass
[Open] Shopee Code League - Logistics,"Logistics Performance
Due to the recent COVID-19 pandemic across the globe, many individuals are increasingly turning to online platforms like Shopee to purchase their daily necessities. This surge in online orders has placed a strain onto Shopee and our logistics providers but customer expectations on the timely delivery of their goods remain high. On-time delivery is arguably one of the most important factors of success in the eCommerce industry and now more than ever, we need to ensure the orders reach our buyers on time in order to build our users confidence in us.
In order to handle the millions of parcels that need to be delivered everyday, we have engaged multiple logistics providers across the region. Only the best logistics providers that are able to meet Shopees delivery standards are partnered with us.
The performance of these providers is monitored regularly and each provider is held accountable based on the Service Level Agreements (SLA). Late deliveries are flagged out and penalties are imposed on the providers to ensure they perform their utmost.
The consistent monitoring and process of holding our logistics providers accountable allows us to maintain our promise of timely deliveries to our buyers.
Task
Identify all the orders that are considered late depending on the Service Level Agreements (SLA) with our Logistics Provider.
For the purpose of this question, assume that all deliveries are considered successful by the second attempt.
Basic Concepts
Each orderid represents a distinct transaction on Shopee.
SLA can vary across each route (A route is defined as Sellers Location to Buyers Location) - Refer to SLA_matrix.xlsx
Pick Up Time is defined as the time when the 3PL picks up the parcel and begins to process for delivery. It marks the start of the SLA calculation.
Delivery Attempt is defined as an attempt made by the 3PL to deliver the parcel to the customer. It may or may not be delivered successfully. In the case when it is unsuccessful, a 2nd attempt will be made. A parcel that has no 2nd attempt is deemed to have been successfully delivered on the 1st attempt.
All time formats are stored in epoch time based on Local Time (GMT+8).
Only consider the date when determining if the order is late; ignore the time.
Working Days are defined as Mon - Sat, Excluding Public Holidays.
SLA calculation begins from the next day after pickup (Day 0 = Day of Pickup; Day 1 = Next Day after Pickup)
2nd Attempt must be no later than 3 working days after the 1st Attempt, regardless of origin to destination route (Day 0 = Day of 1st Attempt; Day 1 = Next Day after 1st Attempt).
Only consider the date when determining if the order is late; ignore the time.
Assume the following Public Holidays: 
2020-03-08 (Sunday);
2020-03-25 (Wednesday);
2020-03-30 (Monday);
2020-03-31 (Tuesday)
Submission Format
Check each delivery order and determine whether it is late.
Two columns required:
orderid.
is_late: assign value 1 if the order is late, otherwise 0.
orderid
is_late
1955512445
0
1955598428
1
Your submission should have 3,176,313 rows (excluding headers), each with 2 columns.
Tips:
1) You are advised to run your tests on a sample of the dataset first.
2) If you are unable to solve the entire problem within the time limit, create the output csv with the required number of columns and rows based on a subset of the problem first.",tabular data,inclass,Logistics Performance,matthewscorrelationcoefficient,[open]-shopee-code-league-logistics,correlation
I-RICH ML COMPETITION,kompetisi yang dibuat untuk mengasah kemampuan anggota KD I-RICH,tabular data,inclass,kompetisi untuk mengasah kemampuan anggota KD I-RICH 2019,mae,i-rich-ml-competition,mae
SERPRO - Iris,"Sobre o conjunto de dados
O conjunto de dados Iris foi usado no clssico artigo de 1936 de R. A. Fisher, The Use of Multiple Measurements in Taxonomic Problems, e tambm pode ser encontrado no UCI Machine Learning Repository.
Ele inclui trs espcies de ris com 50 amostras cada, bem como algumas propriedades sobre cada flor. Uma espcie de flor  linearmente separvel das outras duas, mas as outras duas no so linearmente separveis umas das outras.
Colunas disponveis
As colunas neste conjunto de dados so:
Id (identificao nica)
SepalLengthCm (comprimento da spala - em centmetros)
SepalWidthCm (largura da spala - em centmetros)
PetalLengthCm (comprimento da ptala - em centmetros)
PetalWidthCm (largura da ptala - em centmetros)
Species (nome da espcie)",tabular data,inclass,Classifique plantas íris em três espécies distintas neste clássico conjunto de dados,categorizationaccuracy,serpro-iris,categorization_accuracy
Property price prediction challenge,"Challenge Description
The aim of this challenge is to predict the prices of properties in Washington DC by exploring the various characteristics of properties and their effect on the sales price.
The dataset provided consists of 47 explanatory variables describing various aspects of residential homes. This dataset contains information on real property sales for sales between May 1947 to July 2018 for properties located in Washington DC.
Practice Skills
Creative feature engineering
Advanced regression techniques like random forest and gradient boosting in addition to ensemble learning and stacking methods.
Acknowledgements
All data is available at Open Data DC. The residential and address point data is managed by the Office of the Chief Technology Officer.
Distribution Liability: data terms and conditions",tabular data,inclass,Challenge yourselves to the property price prediction challenge hosted by GA technologies,rmsle,property-price-prediction-challenge,percentage_errors
SERPRO - Abalone,"O que  abalone?
Abalone  um gnero de moluscos gastrpodes marinhos da famlia Haliotidae e o nico gnero catalogado desta famlia. Contm diversas espcies em guas costeiras de quase todo o mundo.
O abalone  muito valorizado na gastronomia de pases asiticos. Alm disso, sua concha  comumente usada na criao de joias, especialmente devido ao seu brilho iridescente.
Suas dimenses variam de dois a trinta centmetros. Devido  sua demanda e alto valor econmico, muitas vezes o abalone  colhido em fazendas e, como tal, existe a necessidade de prever a sua idade.
Como  calculada a idade de um abalone?
A abordagem tradicional para determinar sua idade  cortando a concha atravs do cone, manchando-a e contando o nmero de anis atravs de um microscpio - uma tarefa tediosa e demorada.
Algumas medidas fsicas, mais fceis de obter, podem ser usadas para prever a idade de um abalone. Outras informaes, como padres climticos e localizao (portanto, disponibilidade de alimentos) podem ser necessrias para solucionar o problema.
Sobre o conjunto de dados
O conjunto de dados em questo pode ser usado para obter um modelo matemtico para prever a idade de um abalone a partir de suas medies fsicas.
Colunas disponveis
As colunas neste conjunto de dados so:
id (Inteiro): identificao nica de cada indivduo
sex (String): gnero do indivduo, pode ser M: macho, F: fmea e I: infantil
length (Real): Comprimento - maior medida da concha (em mm)
diameter (Real): Medida perpendicular ao comprimento (em mm)
height (Real): Altura com a carne na concha (em mm)
whole_weight (Real): Peso do abalone inteiro (em g)
shucked_weight (Real): Peso exclusivo da carne (em g)
viscera_weight (Real): Peso das vsceras aps secagem (em g)
shell_weight (Real): Peso da concha depois de seca (em g)
rings (Inteiro): Quantidade de anis (equivalente  idade em anos)",tabular data,inclass,Previsão da idade de moluscos a partir de medições físicas,rmse,serpro-abalone,mse
2019 SMHRD  (  ),     $50k   .,tabular data,inclass,주택 가격 예측 경진 대회(Regression),categorizationaccuracy,2019-smhrd--(--),categorization_accuracy
Competencia-Series-Temporales,"Tamao Dataset: 2075259 (Se ver reducido para la evaluacin en Kaggle)
Caracterstica del DataSet: Multivariate, Time-Series
Justificacin de recoleccin de datos: Predecir la cantidad global activa de consumo de energa promedio por minuto
Consideraciones: El indicador a evaluar ser RMSE.",tabular data,inclass,Esta competencia será sobre el problema de poder predecir el uso de Consumo de Energía en una casa a través de las horas. ,rmse,competencia-series-temporales,mse
Predice el futuro,"Statistical forecasting is the art and science of forecasting from data, with or without knowing in advance what equation you should use. The idea is simple: look for statistical patterns in currently available data that you believe will continue into the future. In other words, figure out the way in which the future will look very much like the present, only longer.
Time Series forecasting techniques are very important for a data scientist to master as time series data occur in every domain from medical, stock market, climate change prediction, and so on.
In that note, here we have presented in front of you an opportunity to master time series forecasting techniques and along with that stand an opportunity to get an internship with TerraBlue XT.",tabular data,inclass,Forecast the values of feature one. ,rmse,predice-el-futuro,mse
Predict the Income - <DECODE> WITH BOARD,"The World Department of Commerce launched Census Bureau to gather data on the countrys earnings, employment, and demographics. Its mission is to become the main source of public data about the nations people and economy. The Census data informs for instance on individuals age, level of education, employment type and income. This information could be hugely significant to businesses. For instance, think of the businesses which target individuals with high income. These businesses could use machine learning models trained on US Census data to predict someones income. If this persons income is higher than a given threshold (for instance $50,000 per year), then the company decides to reach out to him. Better detection and segmentation of potential customers reduces marketing costs, increases conversion rate and thus improves return on investment
Overview of How Kaggles Competitions Work
1. Join the Competition
Read about the challenge description, accept the Competition Rules and gain access to the competition dataset.
2. Get to Work
Download the data, build models on it locally or on Kaggle Kernels (our no-setup, customizable Jupyter Notebooks environment with free GPUs) and generate a prediction file.
3. Make a Submission
Upload your prediction as a submission on Kaggle and receive an the score based on the metric(here its F1 Score).
4. Check the Leaderboard
See how your model ranks against other Kagglers on our leaderboard.
5. Final Evaluation
The top 5 teams on the private leaderboard will be selected based on their metric score. Then out of 10 teams, the team with best Kernel will win. Visit Evaluations tab for more info.
How to Submit your Prediction to Kaggle
Once youre ready to make a submission and get on the leaderboard: 1. Click on the Submit Predictions button 2. Upload a CSV file in the submission file format. Youre able to submit 15 submissions a day.
Submission File Format:
You should submit a csv file with exactly 8561 entries plus a header row. Your submission will show an error if you have extra columns (beyond ID and Income) or rows.
The file should have exactly 2 columns:
ID (sorted in increasing order)
Income (contains your predictions)
Got it! Im ready to get started.
Where do I get help if I need it?
Technical Help: Kaggle Contact Us Page or else react out to Sudip
We encourage you to use the forums often. If you share your knowledge, youll find that others will share a lot in turn!",tabular data,inclass,Predict the income of individual belonging from various demographics,meanfscore,predict-the-income-<decode>-with-board,f_score_multiclass
Seleksi Calon Asisten GAIB,"Background
PT. GAIB FINTECH NUSANTARA is a Fintech company with an app that helps users conduct digital financial transactions. To impress their investors they want to develop a machine learning based model to predict when their users are going to churn so they can do targeted marketing. Customer churn is a condition where a customer stops using a companys product or service after a certain period of time. For this case, the period of churn is set to one month of inactivity (0 transactions in one month).
Task
As the bright new Data Scientist in the company, you are tasked to help them build this model from scratch. You are given a dataset from the data analyst team containing user activity data. The data is activity done in one month starting from the date collected attribute. Each data is given the label isChurned which indicates wether or not that user churns in the next month .
Illustration
user_id date_collected total_transaction isChurned
1 2019-01-01 100000 1
2 2019-02-01 75000000 0
Informasi yang didapat :
User 1 melakukan total transaksi senilai Rp.100.000 dari tanggal 1 Januari 2019 sampai tanggal 31 Januari 2019.
User 1 melakukan churn pada bulan selanjutnya (1 Februari - 28 Februari)
User 2 melakukan total transaksi senilai Rp.75.000.000 dari tanggal 1 Februari 2019 sampai tanggal 28 Februari 2019.
User 2 TIDAK melakukan churn pada bulan selanjutnya
Important : DO NOT SHARE COMPETITION LINK OUTSIDE OF GAIB LAB SELECTION PARTICIPANTS",tabular data,inclass,Seleksi Gaib Tugas III,meanfscore,seleksi-calon-asisten-gaib,f_score_multiclass
Summer Analytics 2020 Capstone Project,"As the COVID-19 keeps unleashing its havoc, the world continues to get pushed into the crisis of the great economic recession, more and more companies start to cut down their underperforming employees. Companies firing hundreds and thousands of Employees is a typical headline today. Cutting down employees or reducing an employee salary is a tough decision to take. It needs to be taken with utmost care as imprecision in the identification of employees whose performance is attriting may lead to sabotaging of both employees career and the companys reputation in the market.
Aim of The Competition
To predict Employee Attrition by the given data about his/her past history.
Acknowledgements
We thank IBM for providing us with the dataset.",tabular data,inclass,Predicting Employee Attrition in the Dawn of Recession,auc,summer-analytics-2020-capstone-project,auc
Similarity Search Project,"Our second and optional task is to find the closest time series in terms of Euclidean distance in the dataset for all the provided 100 queries of both synthetic dataset and seismic dataset, i.e., similarity search.
Our purpose here is to utilize a summarization method to perform similarity search faster. The observation is that since the summarizations are shorter than the original time series, calculating Euclidean distances using the summarizations is much faster than calculating Euclidean distances between the original time series.
Some time series in the dataset might be skipped without checking its Euclidean distance with the query. This is the case when we say this time series is pruned. The question here is whether utilizing our summarization is safe: Are the series we prune correctly pruned?
Notebook
Submissions have to be in Python
Go to Notebooks, then Your Work, then Create New Notebook (or duplicate the template notebook)
The submitted notebook should include the summarization/reconstruction functions you submitted for the summarization task of the project.
Most importantly, the submitted notebook must include the function similarity that returns the pruning ratio averaged over all the 100 queries from both datasets.
A solution template can be found here https://www.kaggle.com/abdumaa/similarity-search-solution-template
Submission
The submission file is a CV file containing a header id, expected and one more line containing the value 1 and the pruning ratio return by the similarity function
Here the official documentation to submit a solution from a notebook
https://www.kaggle.com/dansbecker/submitting-from-a-kernel
I would rather recommend this youtube video of how to participate and submit a solution in a kaggle competition from a notebook (kernel)https://www.youtube.com/watch?v=GJBOMWpLpTQ
To follow up on the youtube video, if you cannot find the button ""Commit"", so you need to click on ""Save Version"", then choose ""Save & Run All (commit)""
Do not forget to share your submission notebook after you have submitted your solution:
click on the button ""share"",
type ""abdu maa"" into the ""collaborator"" search box
click ""save""",tabular data,inclass,This is the second and optional task of the Data Science Project,rmse,similarity-search-project,mse
Predição de Churn,Prever comportamento para reter clientes.,tabular data,inclass,Predição do número de indivíduos que saem de um grupo coletivo durante um período específico.,auc,opt-churn-2019-1,auc
Предсказание положения космических объектов,"Прогнозирование положения спутников - одна из важнейших задач астрономии. Например, информация о точном положении спутников на орбите необходима, чтобы избежать чрезвычайно опасных столкновений спутников. Каждое столкновение приводит не только к разрушению спутников, но также приводит к тысячам осколков космического мусора. Кроме того, более точный прогноз положения спутника поможет рассчитать более эффективные маневры для экономии топлива и продления срока службы спутника на орбите. Задача - предсказать ошибку между реальным положением спутника в пространстве и данными из упрощенной симуляции.Ошибка рассчитывается по следующей формуле: df[error]  = np.linalg.norm(df[[x, y, z]].values - df[[x_sim, y_sim, z_sim]].values, axis=1)",tabular data,inclass,"Предсказать положение спутников, используя данные симуляции",smape,sputnik,percentage_errors
2019  ML competition with KISTI,"Introduction 본 대회는 한국과학기술정보연구원(KISTI)의 ""2019 과학기술 빅데이터 분석가 과정"" 에서 진행하는 데이터 사이언스 대회입니다. Academic 목적이며, 대한민국 누구나 참여하실 수 있습니다. Competition background RMS 타이타닉의 침몰은 역사상 가장 악명 높은 참사 중 하나입니다. 1912 년 4 월 15 일, 첫 항해 도중 타이타닉은 빙산과 충돌 한 후 침몰하였고, 이로 인해 2224 명의 승객과 승무원 중 1502 명이 사망했습니다. 이 비극은 국제 사회에 큰 충격을 주었고, 선박 안전 규정을 개선하는 계기가 되었습니다.
많은 사망자가 생긴 이유 중 하나는 승객과 승무원을위한 구명정이 충분하지 않았기 때문입니다. 침몰에서 살아남는 데는 여러 요소가 있었겠지만, 여성, 어린이 및 상류층과 같은 특정 그룹의 사람들이 생존 가능성이 더 컸습니다.
이 컴퍼티션에서 우리는 어떤 부류의 사람들이 생존 할 가능성이 높았던 지에 대해 분석을 하고, 이를 기반으로 하여 머신러닝 모델을 만든 뒤 승선한 사람들의 생존유무를 예측합니다.",tabular data,inclass,KISTI 와 함께하는 ML 대회 - 타이타닉 생존자를 예측하라!,accuracy,2019-ml-competition-with-kisti,accuracy
DS特論2019年度 演習課題2,"Steamのゲームの購入とプレイ履歴データです。 取得元: https://www.kaggle.com/tamber/steam-video-games gameをgroup1とgroup2に分けています。userもtrainとtestの2グループに分けていて、group1のプレイ履歴は全userのデータがあります。group2については、trainと分類されたuserのデータしかありません。  （仮想シナリオとして、group2が新作ゲームで、trainのuserは先行プレイができた、ということにしましょう。実際にはランダムに分けているので、例えばgroup2に何かの第1作があって、group1にその続編がある、なんてことがあるかも。。。）
testに分類されたuserがgroup2のどのゲームをpurchaseするか予測してください。",tabular data,inclass, Recommendation (steam-video-games),map@K,ds2019uec-task2,reconstruction
Competição DSA de Machine Learning,"Seja bem-vindo à Edição de Janeiro/2019 da Competição DSA de Machine Learning. Esta é uma excelente oportunidade para você praticar tudo que aprendeu nos cursos da Data Science Academy, desenvolver suas habilidades de solução de problemas e ainda ganhar prêmios. Nesta competição seu trabalho será criar um modelo de Machine Learning para prever a ocorrência de doenças.",tabular data,inclass,Competição DSA de Machine Learning - Edição Janeiro/2019,cateagorizationaccuracy,competicao-dsa-machine-learning-jan-2019,categorization_accuracy
Классификация компьютерных атак,"При экспоненциальном росте размеров компьютерных сетей и разработанных приложений становится очевидным значительное увеличение потенциального ущерба, который может быть вызван компьютерными атаками. В этом соревновании представлен надежный набор данных, который содержит нормальные сетевые потоки и 14 классов сетевых потоков компьютерных атак, которые соответствуют критериям реального мира и являются общедоступными. Участникам соревнования предлагается использовать алгоритмы машинного обучения, чтобы провести исследование наилучшего набора анализируемых свойств сетевых потоков для обнаружения определенных категорий атак.",tabular data,inclass,Обучить классификатор сетевой статистики NetFlow (один класс нормы и 14 классов сетевых атак с большим дисбалансом выборки).,"
meanfscore",ids-cic-test,f_score
[DM&amp;PR WS18/19] Machine learning competition,"This dataset contains information on default payments, demographic factors, credit data, history of payment, and bill statements of credit card clients in Taiwan from April 2005 to September 2005.The Goal is to create a ML model that predicts if a client will default on credit card payment in next month.",tabular data,inclass,"A private ML competition for DM&PR students, Uni Tübingen",auc,dm-and-pr-ws1819-machine-learning-competition,auc
[DM&amp;PR WS19/20] Machine learning competition,"This dataset contains information on default payments, demographic factors, credit data, history of payment, and bill statements of credit card clients in Taiwan from April 2005 to September 2005.The Goal is to create an ML model that predicts if a client will default on credit card payment in next month.",tabular data,inclass,"A private ML competition for DM&PR students, Uni Tübingen",auc,dm-and-pr-ws1920-machine-learning-competition,auc
Adult Income,Predict the income class of a person,tabular data,inclass,Adult Income Classification,auc,adult-income,auc
Avito Demand Prediction Challenge,"When selling used goods online, a combination of tiny, nuanced details in a product description can make a big difference in drumming up interest. Details like:

And, even with an optimized product listing, demand for a product may simply not existfrustrating sellers who may have over-invested in marketing.
Avito, Russias largest classified advertisements website, is deeply familiar with this problem. Sellers on their platform sometimes feel frustrated with both too little demand (indicating something is wrong with the product or the product listing) or too much demand (indicating a hot item with a good description was underpriced).
In their fourth Kaggle competition, Avito is challenging you to predict demand for an online advertisement based on its full description (title, description, images, etc.), its context (geographically where it was posted, similar ads already posted) and historical demand for similar ads in similar contexts. With this information, Avito can inform sellers on how to best optimize their listing and provide some indication of how much interest they should realistically expect to receive.",tabular data_image data,featured,Predict demand for an online classified ad,RMSE,avito-demand-prediction-challenge,mse
Indoor Location & Navigation,"Your smartphone goes everywhere with youwhether driving to the grocery store or shopping for holiday gifts. With your permission, apps can use your location to provide contextual information. You might get driving directions, find a store, or receive alerts for nearby promotions. These handy features are enabled by GPS, which requires outdoor exposure for the best accuracy. Yet, there are many times when youre inside large structures, such as a shopping mall or event center. Accurate indoor positioning, based on public sensors and user permission, allows for a great location-based experience even when you arent outside.
Current positioning solutions have poor accuracy, particularly in multi-level buildings, or generalize poorly to small datasets. Additionally, GPS was built for a time before smartphones. Todays use cases often require more granularity than is typically available indoors.
In this competition, your task is to predict the indoor position of smartphones based on real-time sensor data, provided by indoor positioning technology company XYZ10 in partnership with Microsoft Research. Youll locate devices using active localization data, which is made available with the cooperation of the user. Unlike passive localization methods (e.g. radar, camera), the data provided for this competition requires explicit user permission. Youll work with a dataset of nearly 30,000 traces from over 200 buildings.
If successful, youll contribute to research with broad-reaching possibilities, including industries like manufacturing, retail, and autonomous devices. With more accurate positioning, existing location-based apps could even be improved. Perhaps youll even see the benefits yourself the next time you hit the mall.
Acknowledgments
XYZ10 is a rising indoor positioning technology company in China. Since 2017, XYZ10 has been accumulating a privacy-sensitive indoor location dataset of WiFi, geomagnetic, and Bluetooth signatures with ground truths from nearly 1,000 buildings.
Microsoft Research is the research subsidiary of Microsoft. Its goal is to advance state-of-the-art computing and solve difficult world research-motivated competition problems through technological innovation in collaboration with academic, government, and industry researchers.",tabular data_image data,research,Identify the position of a smartphone in a shopping mall,IndoorLocalization,indoor-location-&-navigation,custom_loss
Otto Group Product Classification Challenge,"Get started on this competition through Kaggle Scripts
The Otto Group is one of the worlds biggest e-commerce companies, with subsidiaries in more than 20 countries, including Crate & Barrel (USA), Otto.de (Germany) and 3 Suisses (France). We are selling millions of products worldwide every day, with several thousand products being added to our product line.
A consistent analysis of the performance of our products is crucial. However, due to our diverse global infrastructure, many identical products get classified differently. Therefore, the quality of our product analysis depends heavily on the ability to accurately cluster similar products. The better the classification, the more insights we can generate about our product range.
For this competition, we have provided a dataset with 93 features for more than 200,000 products. The objective is to build a predictive model which is able to distinguish between our main product categories. The winning models will be open sourced.",tabular data_image data,featured,Classify products into the correct category,MulticlassLoss,otto-group-product-classification-challenge,multiclass_log_loss
DonorsChoose.org Application Screening,"Founded in 2000 by a high school teacher in the Bronx, DonorsChoose.org empowers public school teachers from across the country to request much-needed materials and experiences for their students. At any given time, there are thousands of classroom requests that can be brought to life with a gift of any amount.
DonorsChoose.org receives hundreds of thousands of project proposals each year for classroom projects in need of funding. Right now, a large number of volunteers is needed to manually screen each submission before its approved to be posted on the DonorsChoose.org website.
Next year, DonorsChoose.org expects to receive close to 500,000 project proposals. As a result, there are three main problems they need to solve:
How to scale current manual processes and resources to screen 500,000 projects so that they can be posted as quickly and as efficiently as possible
How to increase the consistency of project vetting across different volunteers to improve the experience for teachers
How to focus volunteer time on the applications that need the most assistance
The goal of the competition is to predict whether or not a DonorsChoose.org project proposal submitted by a teacher will be approved, using the text of project descriptions as well as additional metadata about the project, teacher, and school. DonorsChoose.org can then use this information to identify projects most likely to need further review before approval.
With an algorithm to pre-screen applications, DonorsChoose.org can auto-approve some applications quickly so that volunteers can spend their time on more nuanced and detailed project vetting processes, including doing more to help teachers develop projects that qualify for specific funding opportunities.
Your machine learning algorithm can help more teachers get funded more quickly, and with less cost to DonorsChoose.org, allowing them to channel even more funding directly to classrooms across the country.
Getting Started with Kernels
Get familiar with the competition data and the machine learning objective quickly using Kernels. Googles engineering education team has put together a starter tutorial implementing benchmark linear classification model.
Acknowledgments
Machine Learning Crash Course was created by Googles engineering education team in partnership with numerous Machine Learning subject matter experts across Google.",tabular data_text data,playground,Predict whether teachers' project proposals are accepted,AUC,donorschoose.org-application-screening,auc
TMDB Box Office Prediction,"Were going to make you an offer you cant refuse: a Kaggle competition!
In a world where movies made an estimated $41.7 billion in 2018, the film industry is more popular than ever. But what movies make the most money at the box office? How much does a director matter? Or the budget? For some movies, its ""You had me at Hello."" For others, the trailer falls short of expectations and you think ""What we have here is a failure to communicate.""
In this competition, youre presented with metadata on over 7,000 past films from The Movie Database to try and predict their overall worldwide box office revenue. Data points provided include cast, crew, plot keywords, budget, posters, release dates, languages, production companies, and countries. You can collect other publicly available data to use in your model predictions, but in the spirit of this competition, use only data that would have been available before a movies release.
Join in, ""make our day"", and then ""youve got to ask yourself one question: Do I feel lucky?""",tabular data_text data,playground,Can you predict a movie's worldwide box office revenue?,RMSLE,tmdb-box-office-prediction,percentage_errors
Outbrain Click Prediction,"The internet is a stimulating treasure trove of possibility. Every day we stumble on news stories relevant to our communities or experience the serendipity of finding an article covering our next travel destination. Outbrain, the webs leading content discovery platform, delivers these moments while we surf our favorite sites.
Currently, Outbrain pairs relevant content with curious readers in about 250 billion personalized recommendations every month across many thousands of sites. In this competition, Kagglers are challenged to predict which pieces of content its global base of users are likely to click on. Improving Outbrains recommendation algorithm will mean more users uncover stories that satisfy their individual tastes.",tabular data_text data,featured,Can you predict which recommended content each user will click?,MAP@{K},outbrain-click-prediction,reconstruction
PetFinder.my Adoption Prediction,"Millions of stray animals suffer on the streets or are euthanized in shelters every day around the world. If homes can be found for them, many precious lives can be saved  and more happy families created.
PetFinder.my has been Malaysias leading animal welfare platform since 2008, with a database of more than 150,000 animals. PetFinder collaborates closely with animal lovers, media, corporations, and global organizations to improve animal welfare.
Animal adoption rates are strongly correlated to the metadata associated with their online profiles, such as descriptive text and photo characteristics. As one example, PetFinder is currently experimenting with a simple AI tool called the Cuteness Meter, which ranks how cute a pet is based on qualities present in their photos.
In this competition you will be developing algorithms to predict the adoptability of pets - specifically, how quickly is a pet adopted? If successful, they will be adapted into AI tools that will guide shelters and rescuers around the world on improving their pet profiles appeal, reducing animal suffering and euthanization.
Top participants may be invited to collaborate on implementing their solutions into AI tools for assessing and improving pet adoption performance, which will benefit global animal welfare.
Important Note
Be aware that this is being run as a Kernels Only Competition, requiring that all submissions be made via a Kernel output.
Photo by Krista Mangulsone on Unsplash",tabular data_text data_image data,featured,How cute is that doggy in the shelter?,QuadraticWeightedKappa,petfinder.my-adoption-prediction,correlation
Bike Sharing Demand,"Get started on this competition through Kaggle Scripts
Bike sharing systems are a means of renting bicycles where the process of obtaining membership, rental, and bike return is automated via a network of kiosk locations throughout a city. Using these systems, people are able rent a bike from a one location and return it to a different place on an as-needed basis. Currently, there are over 500 bike-sharing programs around the world.
The data generated by these systems makes them attractive for researchers because the duration of travel, departure location, arrival location, and time elapsed is explicitly recorded. Bike sharing systems therefore function as a sensor network, which can be used for studying mobility in a city. In this competition, participants are asked to combine historical usage patterns with weather data in order to forecast bike rental demand in the Capital Bikeshare program in Washington, D.C.
Acknowledgements
Kaggle is hosting this competition for the machine learning community to use for fun and practice. This dataset was provided by Hadi Fanaee Tork using data from Capital Bikeshare. We also thank the UCI machine learning repository for hosting the dataset. If you use the problem in publication, please cite:
Fanaee-T, Hadi, and Gama, Joao, Event labeling combining ensemble detectors and background knowledge, Progress in Artificial Intelligence (2013): pp. 1-15, Springer Berlin Heidelberg.",tabular data_time series,playground,Forecast use of a city bikeshare system,RMSLE,bike-sharing-demand,percentage_errors
Sentiment Analysis on Movie Reviews,"""Theres a thin line between likably old-fashioned and fuddy-duddy, and The Count of Monte Cristo ... never quite settles on either side.""
The Rotten Tomatoes movie review dataset is a corpus of movie reviews used for sentiment analysis, originally collected by Pang and Lee [1]. In their work on sentiment treebanks, Socher et al. [2] used Amazons Mechanical Turk to create fine-grained labels for all parsed phrases in the corpus. This competition presents a chance to benchmark your sentiment-analysis ideas on the Rotten Tomatoes dataset. You are asked to label phrases on a scale of five values: negative, somewhat negative, neutral, somewhat positive, positive. Obstacles like sentence negation, sarcasm, terseness, language ambiguity, and many others make this task very challenging.
Kaggle is hosting this competition for the machine learning community to use for fun and practice. This competition was inspired by the work of Socher et al [2]. We encourage participants to explore the accompanying (and dare we say, fantastic) website that accompanies the paper:
http://nlp.stanford.edu/sentiment/
There you will find have source code, a live demo, and even an online interface to help train the model.
[1] Pang and L. Lee. 2005. Seeing stars: Exploiting class relationships for sentiment categorization with respect to rating scales. In ACL, pages 115124.
[2] Recursive Deep Models for Semantic Compositionality Over a Sentiment Treebank, Richard Socher, Alex Perelygin, Jean Wu, Jason Chuang, Chris Manning, Andrew Ng and Chris Potts. Conference on Empirical Methods in Natural Language Processing (EMNLP 2013).",tabular data_time series,playground,Classify the sentiment of sentences from the Rotten Tomatoes dataset,CategorizationAccuracy,sentiment-analysis-on-movie-reviews,categorization_accuracy
Gendered Pronoun Resolution,"Can you help end gender bias in pronoun resolution?
Pronoun resolution is part of coreference resolution, the task of pairing an expression to its referring entity. This is an important task for natural language understanding, and the resolution of ambiguous pronouns is a longstanding challenge.
Unfortunately, recent studies have suggested gender bias among state-of-the-art coreference resolvers. Google AI Language aims to improve gender-fairness in modeling by releasing the Gendered Ambiguous Pronouns (GAP) dataset, containing gender-balanced pronouns (50% of its examples containing feminine pronouns, and 50% containing masculine pronouns).
In this two-stage competition, Kagglers are challenged to build pronoun resolution systems that perform equally well regardless of pronoun gender. Stage twos final evaluation will use a new dataset following the same format. To encourage gender-fair modeling, the ratio of masculine to feminine examples in the official test data will not be known ahead of time.
----------
Please cite the original paper if you use GAP in your work:
@inproceedings{webster2018gap,
  title =     {Mind the GAP: A Balanced Corpus of Gendered Ambiguous Pronouns},
  author =    {Webster, Kellie and Recasens, Marta and Axelrod, Vera and Baldridge, Jason},
  booktitle = {Transactions of the ACL},
  year =      {2018},
  pages =     {to appear},
}",tabular data_time series,research,Pair pronouns to their correct entities,MulticlassLoss,gendered-pronoun-resolution,multiclass_log_loss
COVID19 Global Forecasting (Week 2),"This week 2 forecasting task is now closed for submissions. Click here to visit the week 3 version, and make a submission there.
This is week 2 of Kaggles COVID19 forecasting series, following the Week 1 competition. This is the 2nd of at least 4 competitions we plan to launch in this series.
Background
The White House Office of Science and Technology Policy (OSTP) pulled together a coalition research groups and companies (including Kaggle) to prepare the COVID-19 Open Research Dataset (CORD-19) to attempt to address key open scientific questions on COVID-19. Those questions are drawn from National Academies of Sciences, Engineering, and Medicines (NASEM) and the World Health Organization (WHO).
The Challenge
Kaggle is launching a companion COVID-19 forecasting challenges to help answer a subset of the NASEM/WHO questions. While the challenge involves forecasting confirmed cases and fatalities between April 1 and April 30 by region, the primary goal isnt only to produce accurate forecasts. Its also to identify factors that appear to impact the transmission rate of COVID-19.
You are encouraged to pull in, curate and share data sources that might be helpful. If you find variables that look like they impact the transmission rate, please share your finding in a notebook.
As the data becomes available, we will update the leaderboard with live results based on data made available from the Johns Hopkins University Center for Systems Science and Engineering (JHU CSSE).
We have received support and guidance from health and policy organizations in launching these challenges. Were hopeful the Kaggle community can make valuable contributions to developing a better understanding of factors that impact the transmission of COVID-19.
Companies and Organizations
There is also a call to action for companies and other organizations: If you have datasets that might be useful, please upload them to Kaggles dataset platform and reference them in this forum thread. That will make them accessible to those participating in this challenge and a resource to the wider scientific community.
Acknowledgements
JHU CSSE for making the data available to the public. The White House OSTP for pulling together the key open questions. The image comes from the Center for Disease Control.
This is a Code Competition. Refer to Code Requirements for details.",tabular data_time series,research,Forecast daily COVID-19 spread in regions around world,MCRMSLE,covid19-global-forecasting-(week-2),percentage_errors_multiclass
AI Academy Intermediate Class Competition 1,"Text documents are one of the richest sources of data for businesses.

Well use a public dataset from the BBC comprised of 2225 articles, each labeled under one of 5 categories: business, entertainment, politics, sport or tech.

The dataset is broken into 1490 records for training and 735 for testing. The goal will be to build a system that can accurately classify previously unseen news articles into the right category.

The competition is evaluated using Accuracy as a metric.
Following blog has good information on how to look at the problem. https://cloud.google.com/blog/products/gcp/problem-solving-with-ml-automatic-document-classification",text data,inclass,News Article Categorization,categorizationaccuracy,ai-academy-intermediate-class-competition-1,categorization_accuracy
"Contradictory, My Dear Watson","""when you have eliminated the impossible, whatever remains, however improbable, must be the truth""
-Sir Arthur Conan Doyle
Our brains process the meaning of a sentence like this rather quickly.
Were able to surmise:
Some things to be true: ""You can find the right answer through the process of elimination.
Others that may have truth: ""Ideas that are improbable are not impossible!""
And some claims are clearly contradictory: ""Things that you have ruled out as impossible are where the truth lies.""
Natural language processing (NLP) has grown increasingly elaborate over the past few years. Machine learning models tackle question answering, text extraction, sentence generation, and many other complex tasks. But, can machines determine the relationships between sentences, or is that still left to humans? If NLP can be applied between sentences, this could have profound implications for fact-checking, identifying fake news, analyzing text, and much more.
The Challenge:
If you have two sentences, there are three ways they could be related: one could entail the other, one could contradict the other, or they could be unrelated. Natural Language Inferencing (NLI) is a popular NLP problem that involves determining how pairs of sentences (consisting of a premise and a hypothesis) are related.
Your task is to create an NLI model that assigns labels of 0, 1, or 2 (corresponding to entailment, neutral, and contradiction) to pairs of premises and hypotheses. To make things more interesting, the train and test set include text in fifteen different languages! You can find more details on the dataset by reviewing the Data page.
Today, the most common approaches to NLI problems include using embeddings and transformers like BERT. In this competition, were providing a starter notebook to try your hand at this problem using the power of Tensor Processing Units (TPUs). TPUs are powerful hardware accelerators specialized in deep learning tasks, including Natural Language Processing. Kaggle provides all users TPU Quota at no cost, which you can use to explore this competition. Check out our TPU documentation and Kaggles YouTube playlist for more information and resources.
Recommended Tutorial
We highly recommend Ana Sofia Uzsoys Tutorial that walks you through creating your very first submission step by step with TPUs and BERT.
This is a great opportunity to flex your NLP muscles and solve an exciting problem!
Disclaimer: The dataset for this competition contains text that may be considered profane, vulgar, or offensive.",text data,getting started,Detecting contradiction and entailment in multilingual text using TPUs,categorizationaccuracy,"contradictory,-my-dear-watson",categorization_accuracy
DL in NLP Spring 2019. Classification,"Your task is to classify news articles into tree categories: news, clickbait and other.
UPD 21.04: added unlabeled data",text data,inclass,Train a clickbait detector,meanfscore,dl-in-nlp-spring-2019.-classification,f_score_multiclass
Tweet Sentiment Analysis,"This competition is for the summer interns at LAS (and any other LAS members who want to participate). The goal is to predict the sentiment of tweets scraped from Twitter.
Awards
The public leader-board is calculated on 40% of the test data. The private leader-board is calculated on the other 60%, and will be revealed after the competition ends. Winners are determined by the private leaderboard
If you can, present your code in a notebook so that the other competitors can see them. If you create a cool visualization in your notebook, we may be able to show off your work to the lab.",text data,inclass,Sentiment analysis of tweets on a theme,auc,tweet-sentiment-analysis,auc
Text classification,"     ,          Avito    .
Baseline ",text data,inclass,Многоклассовая классификация текстов объявлений,categorizationaccuracy,text-classification,categorization_accuracy
Data Mining Lab2,"Description
Emotions are expressed in nuanced ways, which varies by collective or individual experiences, knowledge, and beliefs. Therefore, to understand emotion, as conveyed through text, a robust mechanism capable of capturing and modeling different linguistic nuances and phenomena is needed.
Thus, in this competition, we provide a dataset which was crawled from Twitter, and we have already labeled the emotion for these tweets by some specific hashtags in the original text. There are 8 classes (or say emotions) in our dataset: anger, anticipation, disgust, fear, sadness, surprise, trust, and joy.
You have to clean the data by doing some pre-processing first. Then, apply feature engineering or any other data mining technique you have or havent learned in the Data Mining course. The final goal is to learn a model that is able to predict the emotion behind each tweet.
Note
More detail about this assignment 2 is on iLMS (link) and GitHub (Lab2). Remember to fill in your team name Here .
Acknowledgements
We thank IDEA Lab for providing this dataset.",text data,inclass,Emotion Recognition on Twitter,meanfscore,data-mining-lab2,f_score_multiclass
DeepNLP HSE Course," mail.ru

 :  main_category .
 :        .
 
       28-        .  ,          .         .
    :
,  ,    ;
,     ,        / ;
,      ,    ;
,       ,          ;
  .
 
  ,      ,    , ,   ,     .
 
        .     mail.ru group
  :
transfer learning;
data augmentation;
metric learning;
pseudo labeling;
multitask learning",text data,inclass,Millions of trash questions with billions of answers,meanfscore,deepnlp-hse-course,f_score_multiclass
Quora Insincere Questions Classification,"An existential problem for any major website today is how to handle toxic and divisive content. Quora wants to tackle this problem head-on to keep their platform a place where users can feel safe sharing their knowledge with the world.
Quora is a platform that empowers people to learn from each other. On Quora, people can ask questions and connect with others who contribute unique insights and quality answers. A key challenge is to weed out insincere questions -- those founded upon false premises, or that intend to make a statement rather than look for helpful answers.
In this competition, Kagglers will develop models that identify and flag insincere questions. To date, Quora has employed both machine learning and manual review to address this problem. With your help, they can develop more scalable methods to detect toxic and misleading content.
Heres your chance to combat online trolls at scale. Help Quora uphold their policy of Be Nice, Be Respectful and continue to be a place for sharing and growing the worlds knowledge.
Important Note
Be aware that this is being run as a Kernels Only Competition, requiring that all submissions be made via a Kernel output. Please read the Kernels FAQ and the data page very carefully to fully understand how this is designed.",text data,featured,Detect toxic content to improve online conversations,FScore_1,quora-insincere-questions-classification,f_score
Quora Question Pairs,"Where else but Quora can a physicist help a chef with a math problem and get cooking tips in return? Quora is a place to gain and share knowledgeabout anything. Its a platform to ask questions and connect with people who contribute unique insights and quality answers. This empowers people to learn from each other and to better understand the world.
Over 100 million people visit Quora every month, so its no surprise that many people ask similarly worded questions. Multiple questions with the same intent can cause seekers to spend more time finding the best answer to their question, and make writers feel they need to answer multiple versions of the same question. Quora values canonical questions because they provide a better experience to active seekers and writers, and offer more value to both of these groups in the long term.
Currently, Quora uses a Random Forest model to identify duplicate questions. In this competition, Kagglers are challenged to tackle this natural language processing problem by applying advanced techniques to classify whether question pairs are duplicates or not. Doing so will make it easier to find high quality answers to questions resulting in an improved experience for Quora writers, seekers, and readers.",text data,featured,Can you identify question pairs that have the same intent?,LogLoss,quora-question-pairs,log_loss
TensorFlow 2.0 Question Answering,"Why is the sky blue?
This is a question an open-domain question answering (QA) system should be able to respond to. QA systems emulate how people look for information by reading the web to return answers to common questions. Machine learning can be used to improve the accuracy of these answers.
Existing natural language models have been focused on extracting answers from a short paragraph rather than reading an entire page of content for proper context. As a result, the responses can be complicated or lengthy. A good answer will be both succinct and relevant.
In this competition, your goal is to predict short and long answer responses to real questions about Wikipedia articles. The dataset is provided by Googles Natural Questions, but contains its own unique private test set. A visualization of examples shows long andwhere availableshort answers. In addition to prizes for the top teams, there is a special set of awards for using TensorFlow 2.0 APIs.
If successful, this challenge will help spur the development of more effective and robust QA systems.
About TensorFlow
TensorFlow is an open source platform for machine learning. With TensorFlow 2.0, tf.keras is the preferred high-level API for TensorFlow, to make model building easier and more intuitive. You may use the tf.keras built-in compile()/fit() methods, or write your own custom training loops. See the Effective TensorFlow 2.0 guide and the tf.keras guide for more details.
TensorFlow 2.0 was recently released and this competition is to challenge Kagglers to use TensorFlow 2.0s APIs focused on usability, and easier, more intuitive development, to make advancements on Question Answering.",text data,featured,Identify the answers to real user questions about Wikipedia page content,NQMicroF1,tensorflow-2.0-question-answering,f_score
Texts classification,"     ,          Avito    .
Baseline ",text data,inclass,Classify texts from VK into 4 categories,categorizationaccuracy,texts-classification,categorization_accuracy
Tweet Sentiment Extraction,"""My ridiculous dog is amazing."" [sentiment: positive]
With all of the tweets circulating every second it is hard to tell whether the sentiment behind a specific tweet will impact a company, or a persons, brand for being viral (positive), or devastate profit because it strikes a negative tone. Capturing sentiment in language is important in these times where decisions and reactions are created and updated in seconds. But, which words actually lead to the sentiment description? In this competition you will need to pick out the part of the tweet (word or phrase) that reflects the sentiment.
Help build your skills in this important area with this broad dataset of tweets. Work on your technique to grab a top spot in this competition. What words in tweets support a positive, negative, or neutral sentiment? How can you help make that determination using machine learning tools?
In this competition weve extracted support phrases from Figure Eights Data for Everyone platform. The dataset is titled Sentiment Analysis: Emotion in Text tweets with existing sentiment labels, used here under creative commons attribution 4.0. international licence. Your objective in this competition is to construct a model that can do the same - look at the labeled sentiment for a given tweet and figure out what word or phrase best supports it.
Disclaimer: The dataset for this competition contains text that may be considered profane, vulgar, or offensive.",text data,featured,Extract support phrases for sentiment labels,Jaccard,tweet-sentiment-extraction,segmentation
What's Cooking?,"Picture yourself strolling through your local, open-air market... What do you see? What do you smell? What will you make for dinner tonight?
If youre in Northern California, youll be walking past the inevitable bushels of leafy greens, spiked with dark purple kale and the bright pinks and yellows of chard. Across the world in South Korea, mounds of bright red kimchi greet you, while the smell of the sea draws your attention to squids squirming nearby. Indias market is perhaps the most colorful, awash in the rich hues and aromas of dozens of spices: turmeric, star anise, poppy seeds, and garam masala as far as the eye can see.
Some of our strongest geographic and cultural associations are tied to a regions local foods. This playground competitions asks you to predict the category of a dishs cuisine given a list of its ingredients. 
Acknowledgements
We want to thank Yummly for providing this unique dataset. Kaggle is hosting this playground competition for fun and practice.",text data,playground,Use recipe ingredients to categorize the cuisine,CategorizationAccuracy,whats-cooking?,categorization_accuracy
Classifying Movie Reviews,"Each example in this classification task is a movie review. The goal is to predict whether the review is a positive or a negative one. The data used for this task is based on the Large Movie Review Dataset v1.0[1].
Each review in this task is characterized by the histogram of the words it contains. You are also provided with the raw text that was used to try out new features.
Andrew L. Maas, Raymond E. Daly, Peter T. Pham, Dan Huang, Andrew
Y. Ng, and Christopher Potts. (2011). Learning Word Vectors for
Sentiment Analysis. The 49th Annual Meeting of the Association
for Computational Linguistics (ACL 2011).",text data,inclass,Rating movies for fun and profit,categorizationaccuracy,classifying-movie-reviews,categorization_accuracy
Hackathon Sentimento,"Virtual Hackathon
Participate in virtual hackathon for scholars of Secure and Private AI Scholarship Challenge from Facebook conducted by #sghackathonorgnizrs. Come join us for a fun filled weekend of coding and competing against each other.
When is it?
Hackathon starts => Saturday 00:01am GMT to Monday 11:59am GMT
Coding Time => Saturday 00:01am GMT to Sunday 11:59pm GMT .
Commiting Kernel => Sunday 00:01am GMT to Monday 11:59am GMT
How to participate?

Use this form to sign up. You can participate alone or as part of a team of up to 4 individuals. Only 1 member of the team needs to fill the form. https://forms.gle/EXVwAntevyexEqYP8. Please join the #sg_hackathon-orgnizrs channel to get the announcements and ask questions.
When will results be announced?
On Wednesday
Acknowledgements
We thank Udacity and Facebook for this opportunity
For more FAQs, please go to our our github page",text data,inclass,Sentiment Analysis,categorizationaccuracy,hackathon-sentimento,categorization_accuracy
Recommender Systems,"Can an automated system recommend a funny joke?
Content
The dataset contains over 1.7 million continuous ratings (-10.00 to +10.00) of 150 jokes from 59,132 users. The ratings are real values ranging from -10.00 to +10.00.",text data,inclass,Lets Compete to Predict a User's Favorite Joke!,categorizationaccuracy,recommender-systems,categorization_accuracy
HTA Tagging,"Blinding (or masking) is the process used in experimental research by which study participants, persons caring for the participants, persons providing the intervention,
data collectors and data analysts are kept unaware of
group assignment (control vs intervention). Blinding
aims to reduce the risk of bias that can be caused by an
awareness of group assignment. With blinding, outcomes can be attributed to the intervention itself and not influenced by behaviour or assessment of outcomes that can result purely from knowledge of group allocation.
Blinding of intervention: the medical treatment method is unknown to the experimenters.
Blinding of outcome assessment: the outcome assessment method is unknown to the experimenters.
This competition is aimed to find the robust NLP technique for correctly classify the label class of the Blinding of intervention and the Blinding of outcome assessment.
Input: Content of Journal Papers
Output: The result classes, can be one of PP, PN, PQ, NP, NN, NQ, QP, QN, QQ which is the concatenation of the Blinding of interventions label class and the Blinding of outcome assessments label class.
Label classes of the blinding of intervention and the Blinding of outcome assessment are defined as follows.
P = Positive, the blinding is performed in the experiment.
N = Negative, the blinding is NOT performed in the experiment.
Q = Question, the blinding is not stated in the paper.",text data,inclass,Classify the medical academic papers based on biases.,categorizationaccuracy,hta-tagging,categorization_accuracy
Bag of Words Meets Bags of Popcorn,"In this tutorial competition, we dig a little ""deeper"" into sentiment analysis. Googles Word2Vec is a deep-learning inspired method that focuses on the meaning of words. Word2Vec attempts to understand meaning and semantic relationships among words. It works in a way that is similar to deep approaches, such as recurrent neural nets or deep neural nets, but is computationally more efficient. This tutorial focuses on Word2Vec for sentiment analysis.
Sentiment analysis is a challenging subject in machine learning. People express their emotions in language that is often obscured by sarcasm, ambiguity, and plays on words, all of which could be very misleading for both humans and computers. Theres another Kaggle competition for movie review sentiment analysis. In this tutorial we explore how Word2Vec can be applied to a similar problem.
Deep learning has been in the news a lot over the past few years, even making it to the front page of the New York Times. These machine learning techniques, inspired by the architecture of the human brain and made possible by recent advances in computing power, have been making waves via breakthrough results in image recognition, speech processing, and natural language tasks. Recently, deep learning approaches won several Kaggle competitions, including a drug discovery task, and cat and dog image recognition.
Tutorial Overview
This tutorial will help you get started with Word2Vec for natural language processing. It has two goals: 
Basic Natural Language Processing: Part 1 of this tutorial is intended for beginners and covers basic natural language processing techniques, which are needed for later parts of the tutorial.
Deep Learning for Text Understanding: In Parts 2 and 3, we delve into how to train a model using Word2Vec and how to use the resulting word vectors for sentiment analysis.
Since deep learning is a rapidly evolving field, large amounts of the work has not yet been published, or exists only as academic papers. Part 3 of the tutorial is more exploratory than prescriptive -- we experiment with several ways of using Word2Vec rather than giving you a recipe for using the output.
To achieve these goals, we rely on an IMDB sentiment analysis data set, which has 100,000 multi-paragraph movie reviews, both positive and negative. 
Acknowledgements
This dataset was collected in association with the following publication:
Andrew L. Maas, Raymond E. Daly, Peter T. Pham, Dan Huang, Andrew Y. Ng, and Christopher Potts. (2011). ""Learning Word Vectors for Sentiment Analysis."" The 49th Annual Meeting of the Association for Computational Linguistics (ACL 2011). (link)
Please email the author of that paper if you use the data for any research applications. The tutorial was developed by Angela Chapman during her summer 2014 internship at Kaggle.",text data,getting started,Use Google's Word2Vec for movie reviews,AUC,bag-of-words-meets-bags-of-popcorn,auc
Google QUEST Q&A Labeling,"Computers are really good at answering questions with single, verifiable answers. But, humans are often still better at answering questions about opinions, recommendations, or personal experiences.
Humans are better at addressing subjective questions that require a deeper, multidimensional understanding of context - something computers arent trained to do wellyet.. Questions can take many forms - some have multi-sentence elaborations, others may be simple curiosity or a fully developed problem. They can have multiple intents, or seek advice and opinions. Some may be helpful and others interesting. Some are simple right or wrong.
Unfortunately, its hard to build better subjective question-answering algorithms because of a lack of data and predictive models. Thats why the CrowdSource team at Google Research, a group dedicated to advancing NLP and other types of ML science via crowdsourcing, has collected data on a number of these quality scoring aspects.
In this competition, youre challenged to use this new dataset to build predictive algorithms for different subjective aspects of question-answering. The question-answer pairs were gathered from nearly 70 different websites, in a ""common-sense"" fashion. Our raters received minimal guidance and training, and relied largely on their subjective interpretation of the prompts. As such, each prompt was crafted in the most intuitive fashion so that raters could simply use their common-sense to complete the task. By lessening our dependency on complicated and opaque rating guidelines, we hope to increase the re-use value of this data set. What you see is what you get!
Demonstrating these subjective labels can be predicted reliably can shine a new light on this research area. Results from this competition will inform the way future intelligent Q&A systems will get built, hopefully contributing to them becoming more human-like.",text data,featured,Improving automated understanding of complex question answer content,MCSpearmanR,google-quest-q&a-labeling,correlation
Home Depot Product Search Relevance,"Shoppers rely on Home Depots product authority to find and buy the latest products and to get timely solutions to their home improvement needs. From installing a new ceiling fan to remodeling an entire kitchen, with the click of a mouse or tap of the screen, customers expect the correct results to their queries  quickly. Speed, accuracy and delivering a frictionless customer experience are essential.
In this competition, Home Depot is asking Kagglers to help them improve their customers shopping experience by developing a model that can accurately predict the relevance of search results.
Search relevancy is an implicit measure Home Depot uses to gauge how quickly they can get customers to the right products. Currently, human raters evaluate the impact of potential changes to their search algorithms, which is a slow and subjective process. By removing or minimizing human input in search relevance evaluation, Home Depot hopes to increase the number of iterations their team can perform on the current search algorithms.",text data,featured,Predict the relevance of search results on homedepot.com,RMSE,home-depot-product-search-relevance,mse
Jigsaw Multilingual Toxic Comment Classification,"It only takes one toxic comment to sour an online discussion. The Conversation AI team, a research initiative founded by Jigsaw and Google, builds technology to protect voices in conversation. A main area of focus is machine learning models that can identify toxicity in online conversations, where toxicity is defined as anything rude, disrespectful or otherwise likely to make someone leave a discussion. If these toxic contributions can be identified, we could have a safer, more collaborative internet.
In the previous 2018 Toxic Comment Classification Challenge, Kagglers built multi-headed models to recognize toxicity and several subtypes of toxicity. In 2019, in the Unintended Bias in Toxicity Classification Challenge, you worked to build toxicity models that operate fairly across a diverse range of conversations. This year, were taking advantage of Kaggles new TPU support and challenging you to build multilingual models with English-only training data.
Jigsaws API, Perspective, serves toxicity models and others in a growing set of languages (see our documentation for the full list). Over the past year, the field has seen impressive multilingual capabilities from the latest model innovations, including few- and zero-shot learning. Were excited to learn whether these results ""translate"" (pun intended!) to toxicity classification. Your training data will be the English data provided for our previous two competitions and your test data will be Wikipedia talk page comments in several different languages.
As our computing resources and modeling capabilities grow, so does our potential to support healthy conversations across the globe. Develop strategies to build effective multilingual models and youll help Conversation AI and the entire industry realize that potential.
Disclaimer: The dataset for this competition contains text that may be considered profane, vulgar, or offensive.
To get started with TPUs:
Read the TPU documentation one-pager
Then jump right into the Getting Started Notebooks for this competition
Quick note: a TPU is a network-connected accelerator and requires a couple extra lines in your code. Flipping the TPU switch in your notebook will not, by itself, accelerate your code.",text data,featured,Use TPUs to identify toxicity comments across multiple languages,AUC,jigsaw-multilingual-toxic-comment-classification,auc
Personalized Medicine: Redefining Cancer Treatment,"A lot has been said during the past several years about how precision medicine and, more concretely, how genetic testing is going to disrupt the way diseases like cancer are treated.
But this is only partially happening due to the huge amount of manual work still required. Memorial Sloan Kettering Cancer Center (MSKCC) launched this competition, accepted by the NIPS 2017 Competition Track,  because we need your help to take personalized medicine to its full potential.
Once sequenced, a cancer tumor can have thousands of genetic mutations. But the challenge is distinguishing the mutations that contribute to tumor growth (drivers) from the neutral mutations (passengers). 
Currently this interpretation of genetic mutations is being done manually. This is a very time-consuming task where a clinical pathologist has to manually review and classify every single genetic mutation based on evidence from text-based clinical literature.
For this competition MSKCC is making available an expert-annotated knowledge base where world-class researchers and oncologists have manually annotated thousands of mutations.
We need your help to develop a Machine Learning algorithm that, using this knowledge base as a baseline, automatically classifies genetic variations.

Kaggle is excited to partner with research groups to push forward the frontier of machine learning. Research competitions make use of Kaggles platform and experience, but are largely organized by the research groups data science team. Any questions or concerns regarding the competition data, quality, or topic will be addressed by them.",text data,research,Predict the effect of Genetic Variants to enable Personalized Medicine,MulticlassLoss,personalized-medicine:-redefining-cancer-treatment,multiclass_log_loss
NTUST: Information Retrieval and Applications,Homework 4: Pseudo-relevance Feedback,text data,inclass,Homework 2: BM25,map@{k},ntust:-information-retrieval-and-applications,reconstruction
Hackathon Sentimento_v2,"Virtual Hackathon
Participate in virtual hackathon for scholars of Secure and Private AI Scholarship Challenge from Facebook conducted by #sghackathonorgnizrs. Come join us for a fun filled 5 day of coding and competing against each other.
When is it?
Hackathon starts => Saturday 00:01am GMT to Thursday 00:01am GMT
Coding Time => Saturday 00:01am GMT to Wednesday 00:01am GMT .
Commiting Kernel => Wednesday 00:01am GMT to Thursday 00:01am GMT
How to participate?

Use this form to sign up. You can participate alone or as part of a team of up to 4 individuals. Only 1 member of the team needs to fill the form. https://forms.gle/EXVwAntevyexEqYP8. Please join the #sg_hackathon-orgnizrs channel to get the announcements and ask questions.
When will results be announced?
On Friday
Acknowledgements
We thank Udacity and Facebook for this opportunity
For more FAQs, please go to our our github page",text data,inclass,Sentiment Analysis with tweets,categorizationaccuracy,hackathon-sentimento_v2,categorization_accuracy
Toxic Comment Classification Challenge,"Discussing things you care about can be difficult. The threat of abuse and harassment online means that many people stop expressing themselves and give up on seeking different opinions. Platforms struggle to effectively facilitate conversations, leading many communities to limit or completely shut down user comments.
The Conversation AI team, a research initiative founded by Jigsaw and Google (both a part of Alphabet) are working on tools to help improve online conversation. One area of focus is the study of negative online behaviors, like toxic comments (i.e. comments that are rude, disrespectful or otherwise likely to make someone leave a discussion). So far theyve built a range of publicly available models served through the Perspective API, including toxicity. But the current models still make errors, and they dont allow users to select which types of toxicity theyre interested in finding (e.g. some platforms may be fine with profanity, but not with other types of toxic content).
In this competition, youre challenged to build a multi-headed model thats capable of detecting different types of of toxicity like threats, obscenity, insults, and identity-based hate better than Perspectives current models. Youll be using a dataset of comments from Wikipedias talk page edits. Improvements to the current model will hopefully help online discussion become more productive and respectful.
Disclaimer: the dataset for this competition contains text that may be considered profane, vulgar, or offensive.",text data,featured,Identify and classify toxic online comments,MCAUC,toxic-comment-classification-challenge,auc_multiclass
Jigsaw Unintended Bias in Toxicity Classification,"Can you help detect toxic comments  and minimize unintended model bias? Thats your challenge in this competition.
The Conversation AI team, a research initiative founded by Jigsaw and Google (both part of Alphabet), builds technology to protect voices in conversation. A main area of focus is machine learning models that can identify toxicity in online conversations, where toxicity is defined as anything rude, disrespectful or otherwise likely to make someone leave a discussion.
Last year, in the Toxic Comment Classification Challenge, you built multi-headed models to recognize toxicity and several subtypes of toxicity. This years competition is a related challenge: building toxicity models that operate fairly across a diverse range of conversations.
Heres the background: When the Conversation AI team first built toxicity models, they found that the models incorrectly learned to associate the names of frequently attacked identities with toxicity. Models predicted a high likelihood of toxicity for comments containing those identities (e.g. ""gay""), even when those comments were not actually toxic (such as ""I am a gay woman""). This happens because training data was pulled from available sources where unfortunately, certain identities are overwhelmingly referred to in offensive ways. Training a model from data with these imbalances risks simply mirroring those biases back to users.
In this competition, youre challenged to build a model that recognizes toxicity and minimizes this type of unintended bias with respect to mentions of identities. Youll be using a dataset labeled for identity mentions and optimizing a metric designed to measure unintended bias. Develop strategies to reduce unintended bias in machine learning models, and youll help the Conversation AI team, and the entire industry, build models that work well for a wide range of conversations.
Disclaimer: The dataset for this competition contains text that may be considered profane, vulgar, or offensive.
Acknowledgments
The Conversation AI team would like to thank Civil Comments for making this dataset available publicly and the Online Hate Index Research Project at D-Lab, University of California, Berkeley, whose labeling survey/instrument informed the dataset labeling. Wed also like to thank everyone who has contributed to Conversation AIs research, especially those who took part in our last competition, the success of which led to the creation of this challenge.
This is a Kernels-only competition. Refer to Kernels Requirements for details.",text data,featured,Detect toxicity across a diverse range of conversations,JigsawBiasAUC,jigsaw-unintended-bias-in-toxicity-classification,significance
Movie Review Sentiment Analysis (Kernels Only),"""Theres a thin line between likably old-fashioned and fuddy-duddy, and The Count of Monte Cristo ... never quite settles on either side.""
The Rotten Tomatoes movie review dataset is a corpus of movie reviews used for sentiment analysis, originally collected by Pang and Lee [1]. In their work on sentiment treebanks, Socher et al. [2] used Amazons Mechanical Turk to create fine-grained labels for all parsed phrases in the corpus. This competition presents a chance to benchmark your sentiment-analysis ideas on the Rotten Tomatoes dataset. You are asked to label phrases on a scale of five values: negative, somewhat negative, neutral, somewhat positive, positive. Obstacles like sentence negation, sarcasm, terseness, language ambiguity, and many others make this task very challenging.
Kaggle is hosting this competition for the machine learning community to use for fun and practice. This competition was inspired by the work of Socher et al [2]. We encourage participants to explore the accompanying (and dare we say, fantastic) website that accompanies the paper:
http://nlp.stanford.edu/sentiment/
There you will find have source code, a live demo, and even an online interface to help train the model.
[1] Pang and L. Lee. 2005. Seeing stars: Exploiting class relationships for sentiment categorization with respect to rating scales. In ACL, pages 115124.
[2] Recursive Deep Models for Semantic Compositionality Over a Sentiment Treebank, Richard Socher, Alex Perelygin, Jean Wu, Jason Chuang, Chris Manning, Andrew Ng and Chris Potts. Conference on Empirical Methods in Natural Language Processing (EMNLP 2013).",text data,playground,Classify the sentiment of sentences from the Rotten Tomatoes dataset,CategorizationAccuracy,movie-review-sentiment-analysis-(kernels-only),categorization_accuracy
Natural Language Processing with Disaster Tweets,"Welcome to one of our ""Getting Started"" competitions 
This particular challenge is perfect for data scientists looking to get started with Natural Language Processing. The competition dataset is not too big, and even if you dont have much personal computing power, you can do all of the work in our free, no-setup, Jupyter Notebooks environment called Kaggle Notebooks.
Competition Description
Twitter has become an important communication channel in times of emergency.
The ubiquitousness of smartphones enables people to announce an emergency theyre observing in real-time. Because of this, more agencies are interested in programatically monitoring Twitter (i.e. disaster relief organizations and news agencies).
But, its not always clear whether a persons words are actually announcing a disaster. Take this example:




The author explicitly uses the word ABLAZE but means it metaphorically. This is clear to a human right away, especially with the visual aid. But its less clear to a machine.
In this competition, youre challenged to build a machine learning model that predicts which Tweets are about real disasters and which ones arent. Youll have access to a dataset of 10,000 tweets that were hand classified. If this is your first time working on an NLP problem, weve created a quick tutorial to get you up and running.
Disclaimer: The dataset for this competition contains text that may be considered profane, vulgar, or offensive.
Acknowledgments
This dataset was created by the company figure-eight and originally shared on their Data For Everyone website here.
Tweet source: https://twitter.com/AnyOtherAnnaK/status/629195955506708480",text data,getting started,Predict which Tweets are about real disasters and which ones are not,meanfscore,natural-language-processing-with-disaster-tweets,f_score_multiclass
Whose line is it anyway?,"South African presidents - who said what?
In this competition, you will need to build a sentence classifier to predict the speaker of a sentence from a South African State of the Nation address.
The challenge
The data for this competition consists of all transcripts of SONA speeches between 1990 - 2018. Approximately 30% of the sentences have been extracted from each speech and saved to a separate testing set. The training set is a csv, with one labelled row per speech. Unlike the testing set, these have not been split to sentence level.
You will need to break the speeches up into individual labelled sentences and train a model to classify the sentences in the test set.
The prize
The top scoring submission will receive a cash prize of R1000.",text data,inclass,"Given a line from a state of the nation address, try to predict which president said it!",categorizationaccuracy,whose-line-is-it-anyway?,categorization_accuracy
Sentiment Analysis in Russian,"The main goal of this inclass competition is to identify sentiments of ~2000 news based on predefined training set of ~8000 news by implementing model using Machine Learning algorithms and techniques.
The main language of news is Russian. However, some English and Kazakh names and titles can also be found.",text data,inclass,"Determine sentiments (positive, negative or neutral) of news in russian language.",meanfscore,sentiment-analysis-in-russian,f_score_multiclass
Sarcasmo,"Este problema desafia voc a desenvolver um modelo que detecte sarcasmo no texto. Foi disponibilizado um dataset com 26709 amostras, sendo 18696 para teste.",text data,inclass,20642 - Aprendizado de Máquina - Universidade de Brasília,auc,sarcasmo,auc
Bad comments,"
      1   NLP.
      multilabel :            ,       .
 pipeline :
 
feature extraction
 
   
 :
gensim
fasttext
Glove
BMEmb
spacy
nltk
sklearn.preprocessing
  ,      .",text data,inclass,Соревнование в рамках Лабораторной работы №1,mcauc,bad-comments,auc_multiclass
AS-bow-2019-2020,"Enonc du TD (Analyse de sentiments)
Le TD consiste  finir limplmentation dun prdicteur naf de la polarit dune critique de film tire de IMDB anglais. Le modle est un modle de rgression logistique traditionnel mais implment avec pytorch.
Le jeu de donnes a t tlcharg depuis le site suivant : http://ai.stanford.edu/~amaas/data/sentiment/
Le TD consiste  complter les trous laisss dans le Kernel Point de dpart:
Forker le kernel Point de dpart
Implmenter le chargement de donnes de test dans le notebook dans la cellule marque  cet effet
Ajouter une mthode run_test()  la classe SentimentAnalyser pour quelle produise un jeu de test stock dans un fichier csv.
Soumettre des tests ainsi gnrs  la comptition
Augmenter la mthode train() de la classe SentimentAnalyser pour diviser le jeu de donnes dentrainement en deux sous parties : entrainement et validation (voir librairie torch_utils.data.random_split). Le corps de la mthode sera modifi pour (1) raliser une valuation sur les donnes de validation  chaque poque (2) sauvegarder au final le modle qui minimise la perte sur ces donnes de validation.
Faire une recherche dhyper-paramtres (nombre dpochs et learning rate).
Le rendu attendu est le notebook de Point de dpart qui aura t complt.
Profitez du framework notebook pour commenter vos rponses si besoin.
Barme
Elments indicatifs de notation :
Qualit du code (1)
Avoir soumis au moins un jeu de test valide dans la comptition (1)
Avoir soumis un jeu de test de rsultat nettement suprieur  la baseline (2)
Avoir soumis un jeu de test de rsultat suprieur au corrig du prof (1)",text data,inclass,Analyse de sentiments par sacs de mots (2019-2020),categorizationaccuracy,as-bow-2019-2020,categorization_accuracy
BU CS506 Spring 2020 Midterm,"The goal of this competition is to predict star rating associated with user reviews from Amazon Movie Reviews using the available features.
You are allowed to use any technique used in class for your predictions, as well as classical machine learning algorithms, like random forests, regression trees, etc. Using deep learning models, or any other related technique, that lies far from the syllabus of this class is prohibited. What we mainly seek from this competition - besides performance- is smart ways to make sense of the data, construct new features from the available metadata, and understand your thought procedure.
In addition to submitting your solution online, you need to provide us with a 2-page writeup that describes the algorithm you have implemented and the special tricks you used in order to make it work (or improve). It is important that you show us your thought procedure. Also, describe your strategy for selecting that particular algorithm and how you did your offline evaluation. Note that some sort of offline evaluation is required.
Your writeup should not exceed 2 pages under any circumstance, else it will not be graded.
Your grade will be determined based on your performance at the Kaggle competition, as well from a report that you need to send us at cs506kaggle@gmail.com.
Please email your Kaggle username to cs506kaggle@gmail.com and clearly mention it on your project writeup as well.
Have fun and good luck!!",text data,inclass,The goal of this competition is to predict star ratings using the Amazon Movie Reviews dataset.,rmse,bu-cs506-spring-2020-midterm,mse
Personality Profile Prediction,"MBTI personality profile prediction
In this challenge, you will be required to build and train a model (or many models) capable of predicting a persons MBTI label using only what they post in online forums.
This challenge will require the use of Natural Language Processing to convert the data into machine learning format. This data will then be used to train a classifier capable of assigning MBTI labels to a persons online forum posts.
Read more about the MBTI personality types here OR, better yet, take the test for yourself!
Each MBTI personality type consists of four binary variables, they are:
Mind: Introverted (I) or Extraverted (E)
Energy: Sensing (S) or Intuitive (N)
Nature: Feeling (F) or Thinking (T)
Tactics: Perceiving (P) or Judging (J)
Each person will have only one of the two categories for each variable above. Combining the four variables gives the final personality type. For example, a person who is Extraverted, Intuitive, Thinking and Judging will get the ENTJ personality type.
You will need to build and train a model that is capable of predicting labels for each of the four MBTI variables - i.e. predict four separate labels for each person which, when combined, results in that persons personality type.",text data,inclass,Classify a person's MBTI personality type using text from what they post online.,meancolumnwiselogloss,personality-profile-prediction,multiclass_log_loss
Fake News e ML,"Essa  uma competio de teste, com o objetivo de ensinar fundamentos de como participar e submeter uma analise na competio; aplicar a tcnica de NLP com Bag Of Word (BOW).
O problema a ser resolvido  a deteco de Fake News, essa atividade assola o mundo cybernetico gerando desinformao e at eventuais mortes. Com Machine Learning e Inteligencias artificial conseguimos aplicar algumas tcnicas para tentar amenizar esse problema e alertar as pessoas do potencial perigo que  acreditar em uma noticia falsa e tambm ao compartilhar ela.",text data,inclass,Detectando Fake News com Machine Learning,categorizationaccuracy,fake-news-e-ml,categorization_accuracy
DMA Kaggle Challenge,"Welcome to the Kaggle Challenge INFO 254/DATA144!
In this lab, you will get a chance to use and apply all the concepts you have learnt so far. This is also a very good opportunity get your feet wet in data science competitions! You will also have a chance to experience first hand the problems data scientists can face while dealing with a dataset.
The dataset we have chosen is yelp review dataset, similar to those you saw in previous labs, but with more columns.
Your task is to predict the rating category (is_good_rating) based on other features.
Feel free to play around with the data as much as you want. Explore it. Create new features. Drop unnecessary features. Slice and dice, aggregate, drill up and drill down.
For this competition, you need to submit your code via Kaggle as well as bCourses. There should be only one submission per team (on both Kaggle and bCourses). The deliverables include:
Write-up : A brief description of the approach you used to solve the problems. Submit this on bCourses. Be brief but include all that is important. What was the data like? How did you explore it? How did you transform it? Which models did you use and why? What were your results? Make sure to include your team members names as well as your Kaggle team name. You can find a template linked on bCourses.
Code : Include any notebooks or scripts you used in your solution. Submit as a zipped file on bCourses.
Kaggle Submission : Submit your results to Kaggle for automatic scoring.
Note that this competition is worth twice as much as the regular labs. You have the opportunity to score 20 points instead of 10.",text data,inclass,Classify ratings in yelp review datasets,categorizationaccuracy,dma-kaggle-challenge,categorization_accuracy
Japanese Review Rating Prediction,"

Yahoo!45Positive12Negative

NotebooksDiscussion",text data,inclass,日本語レビューデータを予測しよう,auc,japanese-review-rating-prediction,auc
Language Identification,"Introduction
Is this homework assignment you will work on a written Language Identification (LI) task, using a database create from Wikipedia paragraphs.
Task description
Fork the character-based RNN Baseline and create a new notebook with an additional contribution to the analysis, optimization or comparative study of the proposed model and task. Include your comments, results, tables, graphs and conclusions in the notebook.
Assigment suggestions
Hyperparameter optimization: study the performance of the model as a function of one of its parameters: the embedding size, RNN hidden size, number of layers, batch size, optimizer, learning rate or other optimizer parameters, number of epochs, 
Other input features: modify the code to use other features as input such as words, character n-grams, character counts, 
Comparative analysis with other DNN architectures: convolutional neural networks, average-pooling, 
Description and comparative analysis with other classical LI systems. You can use an existing implementation, but you will have to describe it with your own words in the notebook.",text data,inclass,Written language identification of Wikipedia paragraphs,categorizationaccuracy,language-identification,categorization_accuracy
Korean Gender Bias Detection,"On the Internet, freedom of speech through anonymity has been considered advantageous in that many people can express their opinions transparently. However, at the same time, there are also negative impacts, such as severely aggressive or insulting threats committed to a specific individual or group.
Recently, in Korean society, it has been publicized that a series of issues were presumed to be caused by target-specific toxic comments. More specifically, after some celebrities appealed mental damage continuously and openly, there were cases leading to a tragic decision. After, hate speech in on-line space has emerged as a critical social issue, a methodology that delicately prevents problematic text has not been immediately carried out by on-line service providers, and the comment system of Entertainment news is currently closed on main news platforms.
Without a doubt, various attempts have been proposed to handle this problem, and the term existence-based detection system is a representative method. The dictionary of publicly available profanity terms for The Korean language was also distributed and some systems are also known to train and deploy the system based on these factors. Notwithstanding the method of identifying hate speech by lexical matching is straightforward, there are limitations. Such terms may be meta-language that mentions only specific content, and profanity terms can appear in a modified format. Even without such terms, bias and sexual offense implicated in the sentence may be aggressive to target figures or other readers.
In this regard, we present the first Korean corpus annotated with gender bias. We collected comments from the Korean entertainment news platform that incorporates a wide range of users in Korea.
We believe that our initial efforts in constructing Korean gender bias dataset that also regards
insulting, along with the detection system, will be supportive for both social good and sociolinguistic studies.",text data,inclass,Identify gender bias in Korean entertainment news comments,f_{beta},korean-gender-bias-detection,f_score
OCRV Test Task,"     .   14 :
 
  - 
 

FAQ -   
 
  -  
  
 - 
SIM-  



FAQ - 
 ,     (  ),    ,       .
      .
     ,   (,   )     (   ).
 ODS  DataSouls   ,    https://datasouls.com/c/nghack2019",text data,inclass,Продемонстрируйте навыки классификации текста,meanfscore,ocrv-test-task,f_score_multiclass
Movie Genre Classification,":        .
!
     2 :
 WORD-2-VEC, FastText, GLOVE.
 9-     LSTM  TRANSFORMER.
========================
:
[news, musical, drama, romance, war, biography, sci-fi, thriller, fantasy, documentary, reality-tv, adventure, mystery, action, sport, horror, comedy, short, western, talk-show, adult, game-show, music, history, crime, family, animation]",text data,inclass,Predict movie genre,categorizationaccuracy,movie-genre-classification,categorization_accuracy
Sentence Relatedness,"Enonc
Ce second TP porte sur la conception dembeddings de phrases et de prdiction de similarits entre couples de phrases apparies.
Il sagira par exemple de prdire que le couple de phrases :
The girl in the blue coverall is painting
The woman is holding the paintbrush next to the artist s easel
dcrit des phrases de contenu smantique similaire, alors que le couple de phrases :
The man is singing heartily and playing the guitar
A bicyclist is holding a bike over his head in a group of people
est un couple de phrases sans lien smantique apparent.
Enonc
Ce second TP porte sur la conception dembeddings de phrases et de prdiction de similarits entre couples de phrases apparies.
Il sagira par exemple de prdire que le couple de phrases :
The girl in the blue coverall is painting
The woman is holding the paintbrush next to the artist s easel
dcrit des phrases de contenu smantique similaire, alors que le couple de phrases :
The man is singing heartily and playing the guitar
A bicyclist is holding a bike over his head in a group of people
est un couple de phrases sans lien smantique apparent.
Le jeu de donnes utilis pour ce TP est driv de SICK-R, qui lui-mme appartient  lensemble SentEval
Chaque ligne de donnes comporte un identifiant numrique, une phrase A, une phrase B et une similarit  prdire, value dans lintervalle [0,1].
Formellement, soit \( ( {\bf x}_1, {\bf x}_2 ) \) un couple de phrases  tester, on prdit y un rel dans lintervalle [0,1] qui reprsente la similarit entre les phrases.
Chaque embedding de phrase est obtenu par un bi-lstm :
$$
\begin{align}
{\bf h}_1 &= {\it bilstm}(x_1^{(1)} \ldots x_n^{(1)} ) \\
{\bf h}_2 &= {\it bilstm}(x_1^{(2)} \ldots x_m^{(2)} )
\end{align}
$$
La prdiction de similarit est ralise par un modle logistique:
$$
P(Y=1 | {\bf x}_1, {\bf x}_2 ) = \frac{exp\left({\bf w}^T
\left[\begin{array}{ll}
{\bf h}_1 \\
{\bf h}_2 \\
\end{array}\right] \right) }{1 + exp\left({\bf w}^T
\left[\begin{array}{ll}
{\bf h}_1 \\
{\bf h}_2 \\
\end{array}
\right]
\right) }
$$
Le fichier de prdictions  construire comportera sur chaque ligne lidentifiant numrique suivi de la prdiction faite par votre systme.
La tche consiste :
Forker le kernel Sick dpart
Ecrire une mthode train() du modle qui utilisera un bi-LSTM
Ecrire une mthode run_test() du modle qui fera la prdiction sur des exemples de test.
Utiliser la librairie torch.text pour batcher les donnes dentrainement. Cela fait, vous pourrez activer le GPU dans votre kernel et observer lacclration du processus dentrainement
Soumettre votre fichier de prdiction  la comptition
Le tutoriel suivant peut servir dexemple pour comprendre le fonctionnement de la librairie torch.text: https://mlexplained.com/2018/02/08/a-comprehensive-tutorial-to-torchtext
Barme indicatif
Qualit du code (1)
Avoir batch le code et obtenu une excution efficace avec le GPU (1)
Avoir soumis un jeu de test de rsultat nettement suprieur  la baseline (2)
Avoir soumis un jeu de test de rsultat suprieur au corrig du prof (1)",text data,inclass,Sentence embeddings and paraphrase detection (AA3 - TP2),correlationcoefficient,sentence-relatedness,correlation
EPAM: Exercise 1 -  Sentiment Analysis,Predict the number of positive and negative reviews using either classification or deep learning algorithms.,text data,inclass,This is IMDB reviews challenge.,"
meanfscore",epam-ml-training-imdb,f_score
Predict Future Sales,"This challenge serves as final project for the ""How to win a data science competition"" Coursera course.
In this competition you will work with a challenging time-series dataset consisting of daily sales data, kindly provided by one of the largest Russian software firms - 1C Company. 
We are asking you to predict total sales for every product and store in the next month. By solving this competition you will be able to apply and enhance your data science skills.",time series,playground,"Final project for ""How to win a data science competition"" Coursera course",rmse,predict-future-sales,mse
Time Series Classification,"         IEEE     2008 .      ,         .
    500      .        .",time series,inclass,Time Series Toy Competition,auc,time-series-classification,auc
Multiple regression for time series data,"EVRY Norge (   Infopulse Ukraine -  )       3D Digital Twin       ,      (     , , ,  ..)       (Digital Twin)          IoT ,   .
     multiple regression for time series data (    )          Valhall.  kernels   end-to-end baseline   jupyter notebook   ,                 .",time series,inclass,Oil production forecast in the Norwegian oil region Valhall,mse,multiple-regression-for-time-series-data,mse
Deepfake Detection Challenge,"This competition is closed for submissions. Participants selected code submissions were re-run by the host on a privately-held test set and the private leaderboard results have been finalized. Late submissions will not be opened, due to an inability to replicate the unique design of this competition.
Deepfake techniques, which present realistic AI-generated videos of people doing and saying fictional things, have the potential to have a significant impact on how people determine the legitimacy of information presented online. These content generation and modification technologies may affect the quality of public discourse and the safeguarding of human rightsespecially given that deepfakes may be used maliciously as a source of misinformation, manipulation, harassment, and persuasion. Identifying manipulated media is a technically demanding and rapidly evolving challenge that requires collaborations across the entire tech industry and beyond.
AWS, Facebook, Microsoft, the Partnership on AIs Media Integrity Steering Committee, and academics have come together to build the Deepfake Detection Challenge (DFDC). The goal of the challenge is to spur researchers around the world to build innovative new technologies that can help detect deepfakes and manipulated media.
Challenge participants must submit their code into a black box environment for testing. Participants will have the option to make their submission open or closed when accepting the prize. Open proposals will be eligible for challenge prizes as long as they abide by the open source licensing terms. Closed proposals will be proprietary and not be eligible to accept the prizes. Regardless of which track is chosen, all submissions will be evaluated in the same way. Results will be shown on the leaderboard.
The PAI Steering Committee has emphasized the need to ensure that all technical efforts incorporate attention to how the resulting code and products based on it can be made as accessible and useful as possible to key frontline defenders of information quality such as journalists and civic leaders around the world. The DFDC results will be a contribution to this effort and building a robust response to the emergent threat deepfakes pose globally.",video,featured,Identify videos with facial or voice manipulations,LogLoss,deepfake-detection-challenge,log_loss
What's Cooking? (Kernels Only),"Picture yourself strolling through your local, open-air market... What do you see? What do you smell? What will you make for dinner tonight?
If youre in Northern California, youll be walking past the inevitable bushels of leafy greens, spiked with dark purple kale and the bright pinks and yellows of chard. Across the world in South Korea, mounds of bright red kimchi greet you, while the smell of the sea draws your attention to squids squirming nearby. Indias market is perhaps the most colorful, awash in the rich hues and aromas of dozens of spices: turmeric, star anise, poppy seeds, and garam masala as far as the eye can see.
Some of our strongest geographic and cultural associations are tied to a regions local foods. This playground competitions asks you to predict the category of a dishs cuisine given a list of its ingredients. 
Acknowledgements
We want to thank Yummly for providing this unique dataset. Kaggle is hosting this playground competition for fun and practice.",text data,playground,Use recipe ingredients to categorize the cuisine,CategorizationAccuracy,whats-cooking?-(kernels-only),categorization_accuracy
Hash Code Archive - Drone Delivery,"This is a synthetic code challenge to sharpen your programming skills. This problem was first released during the 2016 qualification round of Googles annual coding competition, Hash Code. Weve re-released it as a Playground Code Competition to help you sharpen your skills. Along with the Photo Slideshow Optimization competition, open for late submissions, you can use it as practice in advance of Hash Code 2021.
The Internet has profoundly changed the way we buy things, but the online shopping of today is likely not the end of that change; the expectations for purchase delivery has gone from a week, to two days, to one day, to same day. What about in just a few hours? With drones, this may be possible, and theyll bring a whole new fleet of problems to solve with data science.
Drones are autonomous, electric vehicles often used to deliver online purchases. Current experiments use flying drones, so theyre never stuck in traffic. As drone technology improves every year, there remains a major issue: how would we manage and coordinate all those drones?
In this competition, you are given a hypothetical fleet of drones, a list of customer orders, and availability of the individual products in warehouses. Can you schedule the drone operations so that the orders are completed as soon as possible?
When flying delivery drones become the norm, scheduling is one of the many problems to be solved. Get a head startand improve your data science skills at the same time.
This is a Code Competition. Refer to Code Requirements for details.
Photo by Ian Usher on Unsplash",text data,playground,Can you help coordinate the drone delivery supply chain?,PostProcessorKernelDesc,hash-code-archive-drone-delivery,points
I’m Something of a Painter Myself,"We recognize the works of artists through their unique style, such as color choices or brush strokes. The “je ne sais quoi” of artists like Claude Monet can now be imitated with algorithms thanks to generative adversarial networks (GANs). In this getting started competition, you will bring that style to your photos or recreate the style from scratch! Computer vision has advanced tremendously in recent years and GANs are now capable of mimicking objects in a very convincing way. But creating museum-worthy masterpieces is thought of to be, well, more art than science. So can (data) science, in the form of GANs, trick classifiers into believing you’ve created a true Monet? That’s the challenge you’ll take on! A GAN consists of at least two neural networks: a generator model and a discriminator model. The generator is a neural network that creates the images. For our competition, you should generate images in the style of Monet. This generator is trained using a discriminator. The two models will work against each other, with the generator trying to trick the discriminator, and the discriminator trying to accurately classify the real vs. generated images. Your task is to build a GAN that generates 7,000 to 10,000 Monet-style images.",image data,getting started,Use GANs to create art - will you be the next Monet?,MiFID,i’m-something-of-a-painter-myself,custom_loss
Word vectors,"Introduction
Is this homework assignment you will create, analyze and evaluate word vectors, using a database created from the Catalan Wikipedia.
Task description
You will have to modify the provided baseline notebooks to create a new notebook with an additional contribution to the analysis, optimization or comparative study of the baseline model and word vectors.
Assignment 1 (word vectors)
Improve CBOW model
Position-dependent Weighting. The standard CBOW model (CBOW Training notebook) sums all the context word vectors with the same weight. Implement and evaluate a weighted sum of the context words with:
a) A fixed scalar weight, e.g, (1,2,2,1) to give more weight to the words that are closer to the predicted central word
b) A trained scalar weight for each position
c) A trained vector weight for each position. Each word vector is element-wise multiplied by the corresponding position-dependent weight and then added with the rest of the weighted word vectors.
d) (Optional) Hyperparameter optimization: study the performance of the model as a function of one of its parameters: the embedding size, batch size, optimizer, learning rate/scheduler, number of epochs, sharing input/output embeddings.
e) (Optional) Implement other methods to obtain word embeddings
Evaluate the performance of the improved Catalan word vectors
a) Implement the WordVector class (Word Vector Analysis notebook) with the most_similar and analogy methods to find closest vectors and analogies using the cosine similarity measure.
b) Intrinsic evaluation: perform an informal evaluation finding good and bad examples of closest words and analogies. You can analyze the behavior of the CBOW word vectors for words with multiple meanings, synonyms and antonyms, word frequency, different types of analogies, bias (gender, race, sexual orientation, etc.).
c) Prediction accuracy: compare the accuracy of the implemented CBOW models in the out-of-domain (el Peridico) test set (prediction of the central word given a context of 2 previous and 2 next words). To obtain your score on the competition test set you have to commit your version of the CBOW Training notebook, wait until the training completes, and then ""Submit to competition"" the obtained file (submission.csv) in the Output section of the notebook.
d) (Optional) Visualize word analogies or word clustering properties
Assignment 2 (language modeling)
Improve the prediction accuracy of the baseline TransformerLayer notebook
Suggestions:
Increase the number of TransformerLayers (2 o more)
Multilayer perceptron (MLP) over the concatenated input vectors
TransformerLayer with multihead attention
Replace the last linear layer and the softmax part of the loss with AdaptiveSoftmax
Sharing input/output embeddings
Hyperparameter optimization: embedding size, batch size, pooling layer (mean, max, first, ), optimizer, learning rate/scheduler, number of epochs, etc.
The report should include the modified source code, a simple schematic drawing of the model, your results and conclusions.
Create a comparative table of the studied models respect to the single-layer transformer baseline, including loss, accuracy, training time, number of parameters and hyperparameters differences with at least 4 new models or hyperparameters.",text data,inclass,Catalan word vectors,categorizationaccuracy,word-vectors,categorization_accuracy
Hackathon SF ML 3,"   ,    .
       .
  RMSLE.
can_buy -     ,
can_promote -   ,
category -  ,
contacts_visible -   ,
date_created -   ,
delivery_available -  ,
description - ,
fields -  ,
id -  ,
images - id ,
location - ,
mortgage_available -  ,
name - ,
payment_available -   ,
price - ,
subcategory - ,
subway - ,",tabular data,inclass,Предсказание цены для объявления,rmsle,hackathon-sf-ml-3,percentage_errors
Car loan default,"Use application data to predict loan default.
Try to create new features.
Be careful, not all features are helpful.",tabular data,inclass,Predict car loan default using application data,auc,car-loan-default,auc
KaggleDays Paris,"Introduction
This is the home page of the KaggleDays Paris competition.
Purpose
You are provided with data about the first seven days of Louis Vuitton products after their launch on www.louisvuitton.com. Your goal is to forecast sales in the next three months separately (you have to make three predictions for one item). You can use product descriptions, sales, social media, website navigation, and image data to help Louis Vuitton to predict and organize new products sales in the coming months.
Sponsors
with the support of:
Organizers",tabular data_image data,inclass,KaggleDays Paris Jan 26th 2019 Competition,rmsle,kaggledays-paris,percentage_errors
LyDzC4rN3f67W7z,,tabular data,inclass,[200227] Pingpong AI Research - 윤승원님,categorizationaccuracy,lydzc4rn3f67w7z,categorization_accuracy
SkillFactory | Final hackathon,"    -!
    .
https://www.youtube.com/watch?v=vIci3C4JkL0&t=38s
 :   ,            !
,   ,    -  .        ,     -.
       AUC-ROC,      50% ,  -   50%.
     Jian Yuanga?    !",image data,inclass,Финальный хакатон с культовой задачей!,auc,skillfactory-|-final-hackathon,auc
ACM Machine Learning (SVNIT),"ACM Machine Learning Contest - 1
This contest is the beginner level machine learning competition. You are free to use any language/library/resources you would want to use. It is specifically aimed at machine learning enthusiasts at SVNIT.
I am a beginner. What to do?
There is a page that reads getting started . It has the basic tutorials on how to get started with Kaggle.
Here is a list of tutorials you would find useful to solve this problem.
Numpy Documentation
Pandas Documentation
Scikit-Learn Documentation
Tutorial on Random Forest Classifier
Acknowledgements
@vaibhavgeek @akshaybusa",tabular data,inclass,This is the basic competition on random forest classifier Level: Beginner Level,auc,acm-machine-learning-(svnit),auc
SYDE 522 (Winter 2020),This is the home page of the SYDE 522 data challenge (Assignment 3) for Winter 2020. The task is 8-class image classification. You can have up to 3 submissions per day.,image data,inclass,Data Challenge for SYDE 522 (Assignment 3),categorizationaccuracy,syde-522-(winter-2020),categorization_accuracy
Avaliação de Carros,"O banco de dados de avaliação de carros foi derivado de um modelo de decisão hierárquica simples, originalmente desenvolvido para a demonstração de DEX, M. Bohanec, V. Rajkovic: sistema especialista para a tomada de decisões. Sistemica 1 (1), pp. 145-157, 1990). O modelo avalia carros de acordo com a seguinte estrutura conceitual: Número de Atributos:Valores dos Atributos:buying (preço de compra) v-high, high, med, low maint (preço da manutenção) v-high, high, med, low doors (número de portas de portas) 2, 3, 4, 5-more persons (capacidade das pessoas) 2, 4, more lug_boot (tamanho do porta malas) small, med, big safety (segurança estimada) low, med, high CLASSES vgood = 1 good = 2 acc = 3
unacc = 4",tabular data,inclass,Multiclasse,log_loss,opt-car-2019-1,log_loss
Хакатон от Кафедры ТПСТС МФТИ (level 1),"В этом соревновании требуется разработать алгоритм обнаружения кораблей на данных, полученных в оптическом диапазоне. Результатом работы алгоритма должны быть координаты пикселей изображения (маска), на котором размещен корабль.
Изображения в наборе данных могут не содержать корабли или иметь разную подстилающую поверхность. Целевые объекты (корабли) в наборе данных могут различаться по размеру и классам.
Таким образом, стоит задача семантической сегментации объектов на изображении.",image data,inclass,Сможете найти морские суда на спутниковых снимках?,dice,almaz-antey-hackathon-l1,segmentation
Einfhrung in Kaggle InClass Competitions</title>,"Diese Seite dient als Sekundärmaterial zur Vorstellung von Kaggle Competitions im Rahmen des Open Education Day 2019. Es handelt sich um eine fiktive Regressionsaufgabe. In train.csv sind einige Funktionswertpaare (xi,yi) gegeben. Die Aufgabe besteht darin, für die x-Werte in test.csv eine Vorhersage zu erstellen. Die Daten sind frei erfunden. Stellen Sie sich vor, die Messwerte seien ein Mass für die Algenverschmutzung in Ihrem Privatteich. Gegeben sind die Messwerte der ersten 20 Sommertage- wie wird die Algenverschmutzung wohl in den nächsten 20 Tagen zunehmen? Machen Sie eine Vorhersage für die Zeitpunkte, die in Xtest gegeben sind!",tabular data,inclass,Zahlenraten mit Statistik: Wie sieht die Funktion aus?,mse,OED19-Regression,mse
House Sales ,"The objective of this competition is to allow the participants to experiment with feature selection, feature engineering and regression models. Target value is the price of house sales. Multiple features are provided; you should select the combination of features (and maybe add more features) to minimize the RMSE score. $$RMSE = \sqrt{\sum{i=1}^{n}{rac{(x{i}^{p}-x_{i}^{t})^{2}}{n}}}$$",tabular data,inclass,Estimate price of house sales,rmse,eic-house-sales,mse
Freesound Audio Tagging 2019,"Here's the background: Some sounds are distinct and instantly recognizable, like a baby’s laugh or the strum of a guitar. Other sounds are difficult to pinpoint. If you close your eyes, could you tell the difference between the sound of a chainsaw and the sound of a blender? Because of the vastness of sounds we experience, no reliable automatic general-purpose audio tagging systems exist. A significant amount of manual effort goes into tasks like annotating sound collections and providing captions for non-speech events in audiovisual content. To tackle this problem, Freesound (an initiative by MTG-UPF that maintains a collaborative database with over 400,000 Creative Commons Licensed sounds) and Google Research’s Machine Perception Team (creators of AudioSet, a large-scale dataset of manually annotated audio events with over 500 classes) have teamed up to develop the dataset for this new competition. To win this competition, Kagglers will develop an algorithm to tag audio data automatically using a diverse vocabulary of 80 categories. If successful, your systems could be used for several applications, ranging from automatic labelling of sound collections to the development of systems that automatically tag video content or recognize sound events happening in real time. Ready to raise your game? Join the competition! Note, this competition is similar in nature to this competition with a new dataset, and multi-class labels.",audio,research,Automatically recognize sounds and apply tags of varying natures,weightedlabelrankingaverageprecision,freesound-audio-tagging-2019,auc
Kobe Bryant Shot Selection,"Using 20 years of data on Kobe's swishes and misses, can you predict which shots will find the bottom of the net? This competition is well suited for practicing classification basics, feature engineering, and time series analysis. Practice got Kobe an eight-figure contract and 5 championship rings. What will it get you?",tabular data,playground,Which shots did Kobe sink?,logloss,kobe-bryant-shot-selection,log_loss
RSNA STR Pulmonary Embolism Detection,"In this competition, you’ll detect and classify PE cases. In particular, you'll use chest CTPA images (grouped together as studies) and your data science skills to enable more accurate identification of PE. If successful, you'll help reduce human delays and errors in detection and treatment.",image data,featured,Classify Pulmonary Embolism cases in chest CT scans,weightedmeancolumnwiselogloss,rsna-str-pulmonary-embolism-detection,custom_loss
Google Landmark Retrieval 2020,"Image retrieval is a fundamental problem in computer vision: given a query image, can you find similar images in a large database? This is especially important for query images containing landmarks, which accounts for a large portion of what people like to photograph. In this competition, the developed models are expected to retrieve relevant database images to a given query image (ie, the model should retrieve database images containing the same landmark as the query).",image data,research,"Given an image, can you find all of the same landmarks in a dataset?",custom_loss,landmark-retrieval-2020,custom_loss
Cassava Disease Classification ,"In this competition, we introduce a dataset of 5 fine-grained cassava leaf disease categories with 9,436 labeled images collected during a regular survey in Uganda, mostly crowdsourced from farmers taking images of their gardens, and annotated by experts at the National Crops Resources Research Institute (NaCRRI) in collaboration with the AI lab in Makarere University, Kampala. The dataset consists of leaf images of the cassava plant, with 9,436 annotated images and 12,595 unlabeled images of cassava leaves. Participants can choose to use the unlabeled images as additional training data. The goal is to learn a model to classify a given image into these 4 disease categories or a 5th category indicating a healthy leaf, using the images in the training data (participants can choose to use the unlabeled images in their training data).",image data,inclass,Classify pictures of cassava leaves into 1 of 4 disease categories (or healthy),categorizationaccuracy,cassava-disease,categorization_accuracy
CareerCon 2019 - Help Navigate Robots,"In this competition, you’ll help robots recognize the floor surface they’re standing on using data collected from Inertial Measurement Units (IMU sensors). We’ve collected IMU sensor data while driving a small mobile robot over different floor surfaces on the university premises. The task is to predict which one of the nine floor types (carpet, tiles, concrete) the robot is on using sensor data such as acceleration and velocity. Succeed and you'll help improve the navigation of robots without assistance across many different surfaces, so they won’t fall down on the job.",tabular data_time series,recruitment,Compete to get your resume in front of our sponsors,categorizationaccuracy,career-con-2019,categorization_accuracy
Two Sigma: Using News to Predict Stock Movements,"Can we use the content of news analytics to predict stock price performance? The ubiquity of data today enables investors at any scale to make better investment decisions. The challenge is ingesting and interpreting the data to determine which data is useful, finding the signal in this sea of information.",time series,featured,Use news analytics to predict stock price performance,custom,two-sigma-financial-news,points
Google Cloud & NCAA® ML Competition 2020-NCAAM,competitors will forecast outcomes of all possible matchups in the 2020 NCAA Division I Men’s and Women’s Basketball Championships. You don't need to participate in the first stage to enter the second. The first stage exists to incentivize model building and provide a means to score predictions.,tabular data,featured,Apply Machine Learning to NCAA® March Madness,logloss,google-cloud-ncaa-march-madness-2020-division-1-mens-tournament,log_loss
empty,,,,,,,
Google Cloud & NCAA® ML Competition 2020-NCAAW,competitors will forecast outcomes of all possible matchups in the 2020 NCAA Division I Men’s and Women’s Basketball Championships. You don't need to participate in the first stage to enter the second. The first stage exists to incentivize model building and provide a means to score predictions.,tabular data,featured,Apply Machine Learning to NCAA® March Madness,logloss,google-cloud-ncaa-march-madness-2020-division-1-womens-tournament,log_loss
Avito Duplicate Ads Detection,"For some sellers, this means posting the same ad several times with slightly altered text or photos taken from different angles. To ensure that buyers can easily find what they're looking for without sifting through dozens of deceptively identical ads, Avito is asking Kagglers to develop a model that can automatically spot duplicate ads. With more accurate duplicate ad detection, Avito will make it much easier for buyers to find and make their next purchase with an honest seller.",image data,featured,Can you detect duplicitous duplicate ads?,auc,avito-duplicate-ads-detection,auc
Flavours of Physics: Finding τ → μμμ,The aim of this year's challenge is to find a phenomenon that is not already known to exist – charged lepton flavour violation – thereby helping to establish 'new physics'. ,tabular data,featured,Identify a rare decay phenomenon,custom,flavours-of-physics,points
RSNA Intracranial Hemorrhage Detection,"In this competition, your challenge is to build an algorithm to detect acute intracranial hemorrhage and its subtypes.",image data,featured,Identify acute intracranial hemorrhage and its subtypes,weightedmeancolumnwiselogloss,rsna-intracranial-hemorrhage-detection,log_loss
2019 SMHRD 경진대회 ( 지능형 ),어떤 사람이 일년에 개인 수입이 $50k 달러 초과인지 예측해보자.,tabular data,inclass,개인 수입 예측 경진 대회(Classification),categorizationaccuracy,smhrd12,categorization_accuracy
Задержка рейса самолета,,tabular data_time series,inclass,"Предскажи, на сколько времени задержали рейс",RMSE,departure-delayed,mse
ClassificationOFShields,In this competition street signs will be classified,tabular data,inclass,HSMAWS18/19 Experiment 2,categorizationaccuracy,classificationofshields,categorization_accuracy
Fieldguide Challenge: Moths & Butterflies,"We are conducting the very first Fieldguide Challenge, a FGVCx challenge focusing on fine-grained classification of Lepidoptera (moths & butterflies). There are an estimated 175,000 species of Lepidoptera (leps), most of which are rarely photographed. Species identification is further complicated by distinct adult and immature stages (caterpillars), and often a lack of concrete visual diagnostics for separation between near-identical species. The goal of the competition is to push state-of-the-art classification of real world data that contains high class imbalance and high infraclass variability. Moths and butterflies are by far the most photographed group of animals outside of birds. Having an effective CV component to identification apps would engage tens of thousands of avid citizen scientists. It would allow us to create effective monitoring programs through promotion of night-lighting for moths, which are sensitive environmental indicators. We estimate that at least 50,000 species of moths and butterflies can be identified by images, and we have citizen science groups throughout the world ready to put their smartphone cameras to use. The Fieldguide Challenge dataset contains over 5000 species, with a combined training and testing set of over 530,000 images that have been collected and verified by multiple authors of Fieldguide. The dataset features visually-similar species, many of which are highly polymorphic, various stages in the species’ lifecycle (larvae vs adult), as well as species that are often misidentified as a moth/butterfly (e.g. caddisflies).",image data,inclass,Improve classification accuracy over a dataset that contains high class imbalance and high heterogeneity,MeanBestErrorAtK,fieldguide-challenge-moths-and-butterflies,custom_loss
Анализ потребительской корзины,"В больших сетях супермаркетов могут возникать задачи сегментирования клиентов и работы с ними в рамках индивидуальных предложений. Если мы знаем, что клиент вряд ли вернется к нам вновь, может, нужно дать ему скидку на следующий раз? Или предложить купить скидочную карту? Таким образом, возникает задача предсказания будущего поведения клиента в момент посещения супермаркета. Мы могли бы делать такое предложение во время покупки. В этом случае, часто нам доступна только одна информация - это чек клиента.",tabular data,inclass,Классификация чеков с покупками в сети супермаркетов,multiclassloss,dmia-sport-2019-fall-competition-1,multiclass_log_loss
[DM&PR WS19/20] Machine learning competition,"This is a private ML competition for students of Data Mining and Probabilistic Reasoning class at the University of Tübingen WS19/20. This dataset contains information on default payments, demographic factors, credit data, history of payment, and bill statements of credit card clients in Taiwan from April 2005 to September 2005. The Goal is to create an ML model that predicts if a client will default on credit card payment in next month.",tabular data,inclass,The Goal is to create an ML model that predicts if a client will default on credit card payment in next month.,auc,dm-and-pr-ws1920-machine-learning-competition,auc
[DM&PR WS18/19] Machine learning competition,"This is a private ML competition for students of Data Mining and Probabilistic Reasoning class at the University of Tübingen WS19/20. This dataset contains information on default payments, demographic factors, credit data, history of payment, and bill statements of credit card clients in Taiwan from April 2005 to September 2005. The Goal is to create an ML model that predicts if a client will default on credit card payment in next month.",tabular data,inclass,"A private ML competition for DM&PR students, Uni Tübingen",auc,dm-and-pr-ws1819-machine-learning-competition,auc
Elo Merchant Category Recommendation,"Imagine being hungry in an unfamiliar part of town and getting restaurant recommendations served up, based on your personal preferences, at just the right moment. The recommendation comes with an attached discount from your credit card provider for a local place around the corner! Right now, Elo, one of the largest payment brands in Brazil, has built partnerships with merchants in order to offer promotions or discounts to cardholders. But do these promotions work for either the consumer or the merchant? Do customers enjoy their experience? Do merchants see repeat business? Personalization is key. Elo has built machine learning models to understand the most important aspects and preferences in their customers’ lifecycle, from food to shopping. But so far none of them is specifically tailored for an individual or profile. This is where you come in. In this competition, Kagglers will develop algorithms to identify and serve the most relevant opportunities to individuals, by uncovering signal in customer loyalty. Your input will improve customers’ lives and help Elo reduce unwanted campaigns, to create the right experience for customers.",tabular data_time series,featured,Help understand customer loyalty,RMSE,elo-merchant-category-recommendation,mse
Google Landmark Recognition Challenge,"Did you ever go through your vacation photos and ask yourself: What is the name of this temple I visited in China? Who created this monument I saw in France? Landmark recognition can help! This technology can predict landmark labels directly from image pixels, to help people better understand and organize their photo collections. Today, a great obstacle to landmark recognition research is the lack of large annotated datasets. In this competition, we present the largest worldwide dataset to date, to foster progress in this problem. This competition challenges Kagglers to build models that recognize the correct landmark (if any) in a dataset of challenging test images. Many Kagglers are familiar with image classification challenges like the ImageNet Large Scale Visual Recognition Challenge (ILSVRC), which aims to recognize 1K general object categories. Landmark recognition is a little different from that: it contains a much larger number of classes (there are a total of 15K classes in this challenge), and the number of training examples per class may not be very large. Landmark recognition is challenging in its own way.",image data,research,Label famous (and not-so-famous) landmarks in images,GoogleGlobalAP,landmark-recognition-challenge,detection
OpenVaccine: COVID-19 mRNA Vaccine Degradation Prediction,"In this competition, we are looking to leverage the data science expertise of the Kaggle community to develop models and design rules for RNA degradation. Your model will predict likely degradation rates at each base of an RNA molecule, trained on a subset of an Eterna dataset comprising over 3000 RNA molecules (which span a panoply of sequences and structures) and their degradation rates at each position. We will then score your models on a second generation of RNA sequences that have just been devised by Eterna players for COVID-19 mRNA vaccines. These final test sequences are currently being synthesized and experimentally characterized at Stanford University in parallel to your modeling efforts -- Nature will score your models!",tabular data_text data,research,Urgent need to bring the COVID-19 vaccine to mass production,MWCRMSE,stanford-covid-vaccine,mse
Humpback Whale Identification Challenge,"In this competition, you’re challenged to build an algorithm to identifying whale species in images. You’ll analyze Happy Whale’s database of over 25,000 images, gathered from research institutions and public contributors. By contributing, you’ll help to open rich fields of understanding for marine mammal population dynamics around the globe.",image data,playground,Can you identify a whale by the picture of its fluke?,MAP@{K},whale-categorization-playground,reconstruction
Airbus Ship Detection Challenge,"Airbus is excited to challenge Kagglers to build a model that detects all ships in satellite images as quickly as possible. Can you find them even in imagery with clouds or haze? Here’s the backstory: Shipping traffic is growing fast. More ships increase the chances of infractions at sea like environmentally devastating ship accidents, piracy, illegal fishing, drug trafficking, and illegal cargo movement. This has compelled many organizations, from environmental protection agencies to insurance companies and national government authorities, to have a closer watch over the open seas. Airbus offers comprehensive maritime monitoring services by building a meaningful solution for wide coverage, fine details, intensive monitoring, premium reactivity and interpretation response. Combining its proprietary-data with highly-trained analysts, they help to support the maritime industry to increase knowledge, anticipate threats, trigger alerts, and improve efficiency at sea. A lot of work has been done over the last 10 years to automatically extract objects from satellite images with significative results but no effective operational effects. Now Airbus is turning to Kagglers to increase the accuracy and speed of automatic ship detection.",image data,featured,Find ships on satellite images as quickly as possible,IntersectionOverUnionObjectSegmentationBeta,airbus-ship-detection,segmentation
Google Cloud & NCAA® ML Competition 2018-Men's,"Google Cloud and NCAA® have teamed up to bring you this year’s version of the Kaggle machine learning competition. Another year, another chance to anticipate the upsets, call the probabilities, and put your bracketology skills to the leaderboard test. Kagglers will join the millions of fans who attempt to forecast the outcomes of March Madness® during this years NCAA Division I Men’s and Women’s Basketball Championships. But unlike most fans, you will pick your bracket using a combination of NCAA’s historical data and your computing power, while the ground truth unfolds on national television. In the first stage of the competition, Kagglers will rely on results of past tournaments to build and test models. We encourage you to post any useful external data as a dataset. In the second stage, competitors will forecast outcomes of all possible match-ups in the 2018 NCAA Division I Men’s and Women’s Basketball Championships. You dont need to participate in the first stage to enter the second. The first stage exists to incentivize model building and provide a means to score predictions. The real competition is forecasting the 2018 results.",tabular data_time series,featured,Apply Machine Learning to NCAA® March Madness®,LogLoss,mens-machine-learning-competition-2018,log_loss
Corporación Favorita Grocery Sales Forecasting,"Brick-and-mortar grocery stores are always in a delicate dance with purchasing and sales forecasting. Predict a little over, and grocers are stuck with overstocked, perishable goods. Guess a little under, and popular items quickly sell out, leaving money on the table and customers fuming. The problem becomes more complex as retailers add new locations with unique needs, new products, ever transitioning seasonal tastes, and unpredictable product marketing. Corporación Favorita, a large Ecuadorian-based grocery retailer, knows this all too well. They operate hundreds of supermarkets, with over 200,000 different products on their shelves. Corporación Favorita has challenged the Kaggle community to build a model that more accurately forecasts product sales. They currently rely on subjective forecasting methods with very little data to back them up and very little automation to execute plans. They’re excited to see how machine learning could better ensure they please customers by having just enough of the right products at the right time.",tabular data_time series,featured,Can you accurately predict sales for a large grocery chain?,NWRMSLE,favorita-grocery-sales-forecasting,percentage_errors_multiclass
Google Cloud & NCAA® ML Competition 2019-Men's,"Google Cloud and NCAA® have teamed up to bring you this year’s version of the Kaggle machine learning competition. Another year, another chance to anticipate the upsets, call the probabilities, and put your bracketology skills to the leaderboard test. Kagglers will join the millions of fans who attempt to forecast the outcomes of March Madness® during this years NCAA Division I Men’s and Women’s Basketball Championships. But unlike most fans, you will pick your bracket using a combination of NCAA’s historical data and your computing power, while the ground truth unfolds on national television. In the first stage of the competition, Kagglers will rely on results of past tournaments to build and test models. We encourage you to post any useful external data as a dataset. In the second stage, competitors will forecast outcomes of all possible match-ups in the 2018 NCAA Division I Men’s and Women’s Basketball Championships. You dont need to participate in the first stage to enter the second. The first stage exists to incentivize model building and provide a means to score predictions. The real competition is forecasting the 2018 results.",tabular data_time series,featured,Apply Machine Learning to NCAA® March Madness®,LogLoss,mens-machine-learning-competition-2019,log_loss
Humpback Whale Identification,"In this competition, you’re challenged to build an algorithm to identifying whale species in images. You’ll analyze Happy Whale’s database of over 25,000 images, gathered from research institutions and public contributors. By contributing, you’ll help to open rich fields of understanding for marine mammal population dynamics around the globe.",image data,featured,Can you identify a whale by its tail?,MAP@{K},humpback-whale-identification,reconstruction
Google Landmark Recognition 2020,"Have you ever gone through your vacation photos and asked yourself: What was the name of that temple I visited in China? or Who created this monument I saw in France? Landmark recognition can help! This technology can predict landmark labels directly from image pixels, to help people better understand and organize their photo collections. This competition challenges Kagglers to build models that recognize the correct landmark (if any) in a dataset of challenging test images.",image data,research,Label famous (and not-so-famous) landmarks in images,GoogleGlobalAP,landmark-recognition-2020,reconstruction
Freesound General-Purpose Audio Tagging Challenge,"You’re challenged to build a general-purpose automatic audio tagging system using a dataset of audio files covering a wide range of real-world environments. Sounds in the dataset include things like musical instruments, human sounds, domestic sounds, and animals from Freesound’s library, annotated using a vocabulary of more than 40 labels from Google’s AudioSet ontology. To succeed in this competition your systems will need to be able to recognize an increased number of sound events of very diverse nature, and to leverage subsets of training data featuring annotations of varying reliability (see Data section for more information).",audio,research,Can you automatically recognize sounds from a wide range of real-world environments?,MAP@{K},freesound-audio-tagging,reconstruction
2018 Data Science Bowl,Teams will create a computer model that can identify a range of nuclei across varied conditions.,image data,featured,Find the nuclei in divergent images to advance medical discovery,IntersectionOverUnionObjectSegmentation,data-science-bowl-2018,segmentation
Facebook V: Predicting Check Ins,"The goal of this competition is to predict which place a person would like to check in to. For the purposes of this competition, Facebook created an artificial world consisting of more than 100,000 places located in a 10 km by 10 km square. For a given set of coordinates, your task is to return a ranked list of the most likely places. Data was fabricated to resemble location signals coming from mobile devices, giving you a flavor of what it takes to work with real data complicated by inaccurate and noisy values. Inconsistent and erroneous location data can disrupt experience for services like Facebook Check In.",tabular data,recruitment,Identify the correct place for check ins,MAP@{K},facebook-v-predicting-check-ins,reconstruction
Statoil/C-CORE Iceberg Classifier Challenge,"In this competition, you’re challenged to build an algorithm that automatically identifies if a remotely sensed target is a ship or iceberg. Improvements made will help drive the costs down for maintaining safe working conditions.",image data,featured,"Ship or iceberg, can you decide from space?",LogLoss,statoil-iceberg-classifier-challenge,log_loss
Invasive Species Monitoring,"In this playground competition, Kagglers are challenged to develop algorithms to more accurately identify whether images of forests and foliage contain invasive hydrangea or not. Techniques from computer vision alongside other current technologies like aerial imaging can make invasive species monitoring cheaper, faster, and more reliable.",image data,playground,Identify images of invasive hydrangea,AUC,invasive-species-monitoring,auc
iMet Collection 2019 - FGVC6,"The Metropolitan Museum of Art in New York, also known as The Met, has a diverse collection of over 1.5M objects of which over 200K have been digitized with imagery. The online cataloguing information is generated by Subject Matter Experts (SME) and includes a wide range of data. These include, but are not limited to: multiple object classifications, artist, title, period, date, medium, culture, size, provenance, geographic location, and other related museum objects within The Met’s collection. While the SME-generated annotations describe the object from an art history perspective, they can also be indirect in describing finer-grained attributes from the museum-goer’s understanding. Adding fine-grained attributes to aid in the visual understanding of the museum objects will enable the ability to search for visually related objects.",image data,research,Recognize artwork attributes from The Metropolitan Museum of Art,MeanFScoreBeta,imet-2019-fgvc6,f_score_multiclass
How Much Did It Rain? II,"In this competition, you are given snapshots of polarimetric radar values and asked to predict the hourly rain gauge total. A word of caution: many of the gauge values in the training dataset are implausible (gauges may get clogged, for example). More details are on the data page.",time series,research,Predict hourly rainfall using data from polarimetric radars,MAE,how-much-did-it-rain-ii,mae
Coupon Purchase Prediction,"Using past purchase and browsing behavior, this competition asks you to predict which coupons a customer will buy in a given period of time. The resulting models will be used to improve Ponpare's recommendation system, so they can make sure their customers don't miss out on their next favorite thing.",tabular data_time series,featured,Predict which coupons a customer will buy,MAP@{K},coupon-purchase-prediction,reconstruction
Santander Product Recommendation,"Under their current system, a small number of Santander’s customers receive many recommendations while many others rarely see any resulting in an uneven customer experience. In their second competition, Santander is challenging Kagglers to predict which products their existing customers will use in the next month based on their past behavior and that of similar customers.",tabular data_time series,featured,Can you pair products with people?,MAP@{K},santander-product-recommendation,reconstruction
Carvana Image Masking Challenge,"In this competition, you’re challenged to develop an algorithm that automatically removes the photo studio background. This will allow Carvana to superimpose cars on a variety of backgrounds. You’ll be analyzing a dataset of photos, covering different vehicles with a wide variety of year, make, and model combinations.",image data,featured,Automatically identify the boundaries of the car in an image,Dice,carvana-image-masking-challenge,segmentation
Spooky Author Identification,"In this year's Halloween playground competition, you're challenged to predict the author of excerpts from horror stories by Edgar Allan Poe, Mary Shelley, and HP Lovecraft. We're encouraging you (with cash prizes!) to share your insights in the competition's discussion forum and code in Kernels. We've designated prizes to reward authors of kernels and discussion threads that are particularly valuable to the community. Click the Prizes tab on this overview page to learn more.",tabular data_text data,playground,Share code and discuss insights to identify horror authors from their writings,MulticlassLoss,spooky-author-identification,multiclass_log_loss
TrackML Particle Tracking Challenge,TrackML Particle Tracking Challenge,tabular data,featured,High Energy Physics particle tracking in CERN detectors,TrackML,trackml-particle-identification,points
NFL 1st and Future - Impact Detection,"In this competition, you’ll develop a computer vision model that automatically detects helmet impacts that occur on the field. Kick off with a dataset of more than one thousand definitive head impacts from thousands of game images, labeled video from the sidelines and end zones, and player tracking data. This information is sourced from the NFL’s Next Gen Stats (NGS) system, which documents the position, speed, acceleration, and orientation for every player on the field during NFL games.",image data,features,Detect helmet impacts in videos of NFL plays,custom netric,nfl-impact-detection,detection
COVID19 Global Forecasting (Week 3),"While the challenge involves forecasting confirmed cases and fatalities between April 1 and April 30 by region, the primary goal isn't only to produce accurate forecasts. It’s also to identify factors that appear to impact the transmission rate of COVID-19.",tabular data,research,Forecast daily COVID-19 spread in regions around world,MCRMSLE,covid19-global-forecasting-week-3,percentage_errors_multiclass
Santa 2019: Revenge of the Accountants,Your task is to schedule the families to Santa's Workshop in a way that minimizes the penalty cost to Santa ,tabular data_time series,playground,Oh what fun it is to revise . . .,SantaWorkshopSchedule2019Revenge,santa-2019-revenge-of-the-accountants,custom_loss
Lyft 3D Object Detection for Autonomous Vehicles,"Self-driving technology presents a rare opportunity to improve the quality of life in many of our communities. Avoidable collisions, single-occupant commuters, and vehicle emissions are choking cities, while infrastructure strains under rapid urban growth. Autonomous vehicles are expected to redefine transportation and unlock a myriad of societal, environmental, and economic benefits. You can apply your data analysis skills in this competition to advance the state of self-driving technology. Lyft, whose mission is to improve people’s lives with the world’s best transportation, is investing in the future of self-driving vehicles. Level 5, their self-driving division, is working on a fleet of autonomous vehicles, and currently has a team of 450+ across Palo Alto, London, and Munich working to build a leading self-driving system (they’re hiring!). Their goal is to democratize access to self-driving technology for hundreds of millions of Lyft passengers. From a technical standpoint, however, the bar to unlock technical research and development on higher-level autonomy functions like perception, prediction, and planning is extremely high. This implies technical R&D on self-driving cars has traditionally been inaccessible to the broader research community. This dataset aims to democratize access to such data, and foster innovation in higher-level autonomy functions for everyone, everywhere. By conducting a competition, we hope to encourage the research community to focus on hard problems in this space—namely, 3D object detection over semantic maps. In this competition, you will build and optimize algorithms based on a large-scale dataset. This dataset features the raw sensor camera inputs as perceived by a fleet of multiple, high-end, autonomous vehicles in a restricted geographic area.",image data,featured,Can you advance the state of the art in 3D object detection?,Lyft3DObjectDetectionAP,3d-object-detection-for-autonomous-vehicles,detection
Flavours of Physics: Finding τ  →  μμμ,The aim of this year's challenge is to find a phenomenon that is not already known to exist – charged lepton flavour violation – thereby helping to establish 'new physics'. ,tabular data,featured,Identify a rare decay phenomenon,CernWeightedAuc,flavours-of-physics,points
Cdiscount’s Image Classification Challenge,"In this challenge you will be building a model that automatically classifies the products based on their images. As a quick tour of Cdiscount.com's website can confirm, one product can have one or several images.",image data,featured,Categorize e-commerce photos,CategorizationAccuracy,cdiscount-image-classification-challenge,categorization_accuracy
Car Classification(Project Vision),"Cars are one of the most sought after luxury today. They are mostly admired for their sleek design and features. In this task, you are given 45 classes(cars) with 100 images each.Train a CNN on the given dataset and submit the predictions on the test set in a csv file. Download the dataset from data tab and read further instructions",image data,inclass,Identify the cars in images using CNNs,CategorizationAccuracy,car-classificationproject-vision,categorization_accuracy
RSNA Pneumonia Detection Challenge,"In this competition, you’re challenged to build an algorithm to detect a visual signal for pneumonia in medical images. Specifically, your algorithm needs to automatically locate lung opacities on chest radiographs.",image data,featured,Can you build an algorithm that automatically detects potential pneumonia cases?,RSNAObjectDetectionAP,rsna-pneumonia-detection-challenge,detection