Release Planning of Mobile Apps Based on User Reviews

Author(s): Lorenzo Villarroel, Gabriele Bavota, Barbara Russo, Rocco Oliveto, Massimiliano Di Penta
Venue: International Conference on Software Engineering
Date: May 2016


This paper investigates the issues related to user reviews for applications in the mobile marketplace. Since popular applications could receive hundreds of user reviews per day, it becomes very difficult to manually parse through them all to find relevant information for what features should be implemented or what bugs should be fixed in the next iteration. In order to automate this process, the authors of this paper have introduced a tool called CLAP (Crowd Listener for releAse Planning), which can categorize reviews based on relevant information (bug reports, feature requests, other), cluster related reviews together (same bugs, same features, etc.), and prioritize the clustered reviews for upcoming releases.

This tool uses machine learning and clustering algorithms to parse through all of the user reviews in an effort to generate a list of feature requests and bug reports that have been most requested by the community. The tool was able to produce results with high accuracy in categorizing reviews on the basis of contained information (86%), in creating meaningful clusters of related reviews (77%), and in recommending features to implement and bugs to fix in the next app release (~72%).