Currently, Android applications are becoming more and more popular but almost all Android developers will face to a large number of issues reportsevery day, and thendevelopers will debug their code basedon theissue reports posted by other users.Therefore, analyzing those issuereports of Android apps hosted on GitHub has huge positive effects on Android API development. In the project, based on issuereports text hosted on GitHub, the textRank algorithmand BERT model in the machine learning is usedtoobtain summarieswith a large number issue report. According to comparison of the performance of the two models, although BERT can be said state-of-the-art in NLP techniques, there is still no steady improvement on TextRank in the quality of summaries andcombining context.