A dataset of Mobile application reviews for classifying reviews into software Engineering's maintenance tasks using data mining techniques
Description
This dataset has been collected from two different sources. The first dataset was taken from [1] and collected by Panichella et al. We obtained this dataset from Dr. Sebastiano Panichella via email. This dataset contains reviews of the AngryBirds, Dropbox, and Evernote app, which were taken from Apple’s App Store, other reviews were taken from Android’s Google Play store such as TripAdvisor, PicsArt, Pinterest and Whatsapp. This dataset consist of with 1390 reviews from all previously mentioned apps and all reviews were classified into four classes related to Software engineering’s maintenance task as follows: 192 reviews as Feature Request (FR), 494 reviews as Problem Discovery (PD), 603 reviews as Information Gaing (IG) and 101 reviews as Information Seeking (IS). We indicate to this dataset as “Pan Dataset”. The second dataset is used in [2] and prepared by Maalej et al. It is available at Hamburg University website on this direct link (https://mast.informatik.uni-hamburg.de/app-review-analysis). The truth dataset contains 3691 reviews from different Google’s apps store and Apple’s app store. We indicate to this dataset as “maalej dataset”. All reviews were classified into four classes related to Software engineering’s maintenance task as follows: 252 reviews as Feature Request (FR), 370 reviews as bug report/Problem Discovery (BR/PD), 607 reviews as User Experience (UE) and 2461 reviews as Rating (RT)