Mobile Application User Review Based Feature Request and Bug Discovery

This study focuses on building a deep learning model that can classify mobile app user reviews in the dataset made available by Walid Maalej during their research called " On the automatic classification of app reviews" back in 2015. The dataset consists of 4400 user reviews extracted from the Google store and Apple store. This dataset contains user reviews belonging to four classes ("Rating", "User Experience", "Feature", "Bug") and it also contains the sentiment, tense, ratings and other important metadata related to each user review. At the moment the study focuses on developing and benchmarking the performance of various deep learning models combined with some of the well-known word embeddings such as word2vec, glove, fastext.