Detection of Covid-19 Fake News text data using Random Forest and Decision tree Classifiers

Abstract: In current days, Fake news has been growing with significant numbers for numerous political, business, and social reasons. Such news has also become ubiquitous strongly in the online world or social network. People/virtual communities can get contaminated quickly by such fake news. Such news related to Coronavirus covid-19 has dramatically influenced the offline community also. So, there is required a great from self-awareness and community for and better understanding covid-19 informational change the fake news that people spread for political or economic purposes amid this the Corona pandemic crisis to spread terror in society, without taking considerate the feelings of the public. In research articles, particularly these days within covid-19 informational change, we are interested in knowing the algorithms of text mining and machine learning that have the ability to handle and distinguish among real and fake news for Coronavirus covid-19. Through applied two machine learning supervised algorithms, i.e., Random forest and decision tree classifiers to detect Coronavirus covid-19 fake news with our model, Count Vectorizer and Document Frequency Vectorizer as feature extraction after making a set of the initial set such as preprocess and normalization of the dataset. Our proposed detection and the algorithm's ability to differentiate and verify the real and fake news for covid-19 depends on the polarity of the corresponding data set. Finally, we achieved 94.49 % accuracy with the Random forest classifier. Also, we achieved 92.07% accuracy with the decision tree classifier; all of its results were great with our model.
Keywords: Coronavirus covid-19, Random Forest, Text Classification, Fake News, Natural Language Processing, binary classification, Machine Learning Algorithms, Decision tree, text mining.