Leveraging Insights from Mobile App Reviews to Support Release Planning and Maintenance

Developers seek to know what the users of the mobile apps they developed are saying about their products and what improvements they want. This helps developers update their apps according to the users' feedback and need. The goal of this thesis is to automate the process of leveraging key insights from the users' reviews to help developers better understand key concerns of large user population and plan future releases of their apps. In this thesis, we present an approach that automates the process of classifying user reviews into five categories according to the information contained within them. This information can range from improvements about the app to reporting bugs along with how users like/dislike the app. We then group related reviews into clusters by utilizing the LDA-based topic modelling and vector-space model. Then, for each generated cluster we extract the overall sentiment, generate hot topics and a short summary.