User Feedback Intelligence: On the Automated Processing of User Feedback

Summary of Artifacts
These artifacts accompany the "User Feedback Intelligence: On the Automated Processing of User Feedback" chapter within the chapter within the "Handbook of Natural Language Processing for Requirements Engineering" book. The book chapter includes a "Use Cases" section where natural language processing (NLP) techniques are applied to user feedback from various sources [1, 2]. The Jupyter Notebooks in this replication package can be used to follow along with the use cases in the chapter.

Author and Article Details
Prof. Dr. Walid Maalej (Universität Hamburg) - walid.maalej@uni-hamburg.de

Volodymyr Biryuk (Universität Hamburg) - volodymyr.biryuk@uni-hamburg.de

Jialiang Wei (IMT Mines Alès) - jialiang.wei@mines-ales.fr

Fabian Panse (Hasso-Plattner-Institut) - fabian.panse@hpi.de

Please cite this work as:

Maalej W, Biryuk V, Wei J, Panse F. of Part, "User Feedback Intelligence: On the Automated Processing of User Feedback" in "Handbook of Natural Language Processing for Requirements Engineering", 1st Edition, Ferrari A, Deshpande G. Eds. Springer Nature Switzerland AG, Cham, Switzerland, 2024, to appear.

Title of Chapter: "User Feedback Intelligence: On the Automated Processing of User Feedback"

Abstract of the Chapter: User feedback is becoming an increasingly important source of information for requirements engineering, user experience design, and software engineering in general. User feedback is nowadays largely avail- able and easily accessible online, in social media, product forums, or app stores. Over the last decade, research has shown that user feedback can help software teams a) to identify, reproduce, and fix defects faster, b) to get inspirations and ideas about new features and product improvements, and c) to better understand how users actually are using the products and what their opinions, priorities, and workarounds for specific features or components are. However, to tap the full potential of feedback, there are two main chal- lenges that need to be solved. First, software vendors must deal with the large quantity of feedback data, which is hardly manageable manually. Second, vendors also must deal with the varying quality of feedback as some items might be written in a slang language, uninformative, repet- itive, or simply wrong. This chapter summarizes and discuss how data mining, machine learning, and natural language processing (NLP) tech- niques can be used to cope with the quantity and quality challenges—for an effective analysis and actionable usage of user feedback in software and requirements engineering.

[1] Stanik, C., Häring, M., Maalej, W.: Classifying multilingual user feedback using traditional machine learning and deep learning. In: 27th IEEE International Requirements Engineering Conference Workshops, RE 2019 Workshops, Jeju Island, Korea (South), September 23-27, 2019. pp. 220–226. IEEE (2019).
[2] Wei, J., Courbis, A., Lambolais, T., Xu, B., Bernard, P., Dray, G.: Zeroshot bilingual app reviews mining with large language models. In: 35th IEEE International Conference on Tools with Artificial Intelligence, ICTAI 2023, Atlanta, GA, USA, November 6-8, 2023. pp. 898–904. IEEE (2023).