You are here
Analysing Feature-Specific User Feedback in App Reviews
App reviews are a rich source of on-line users’ feedback on mobile applications containing information such bug reports, feature request or the overall user experience with apps. Existing work proposed several approaches to analyze app reviews. However, this work focuses on the mining techniques with only superficial attention to their software engineering use cases. My thesis will address this concern by providing a unified description of the many potential use cases of mining app reviews and by evaluating such techniques with respect to these use cases. In the seminar, I will present my ongoing work on replicating and expanding previous evaluations of state-of-the-art approaches for feature-based sentiment analysis. I will discuss how we evaluate the accuracy of these approaches in identifying feature expressions and analysing their associated user sentiment. Then, I will discuss a preliminary study on adopting information retrieval techniques to support software engineers in eliciting requirements from app reviews.