Identifying the BUG reports quickly and accurately will help developers to repair mobile applications.Therefore,we propose an automatic identification method of mobile application bug reports based on deep learn-ing by investigating the different stages of mobile application operation and taking user reviews and issue reports as the starting points.Firstly,Word2Vec is used to get word vectors;secondly,a Bi-LSTM network is construc-ted to obtain high-level text features.Furthermore,we assign the weight through the attention mechanism to cap-ture the information in the sentence that plays a key role in the identification of the bug reports.Finally we get the bug label of the issue report.Our method is better than the methods proposed in previous works in terms of rate,and this study expounds the impact of datasets from different perspectives on mobile application BUG report identification,which helps developers to identify mobile application BUG reports accurately and improve the effi-ciency of BUG repair.
关键词
移动应用/软件维护/问题报告/用户反馈/BUG标签
Key words
mobile application/software maintenance/issue report/user review/BUG label
维普
万方数据