Effective Bug Triage Based on a Hybrid Neural Network

With the increasing scale and complexity of open source software, the quality of software has become a focus to which repairers pay close attention. Due to the inevitable existence of some known or unknown bugs in software,under certain conditions, software bugs may directly cause program running errors, and then produce abnormal running results and wrong program behavior, which will cause huge economic losses. Therefore, software defect repair is an important part of software evolution and quality assurance. Quickly and efficiently assigning defect reports to the right repairer for repair,to ensure efficiency and reduce the cost of open-source software development is an important problem that must be solved in software quality improvement. In this study, we propose a new defect report repair recommendation algorithm, RCNN, which can effectively learn the features of the defect report and recommend the appropriate repairer according to the feature. The proposed algorithm uses a CNN convolution kernel to capture the local information of the text and RNN is used to capture the sequence information of the text. The attention mechanism is introduced to learn the contribution ratio of each part of the text to the overall semantic information of the text. Thus, to a certain extent, it makes up for the defect that RNN cannot effectively learn and monitor remote information. Through experiments on the Eclipse and Mozilla datasets, compared with NB (naive Bayes), SVM (support vector machines), LeeCNN and DBRNNA, the RCNN model can effectively find the appropriate bug repairer among many repairers, and achieve higher classification accuracy.