Security and performance bug reports are high impact bugs of software. So, these type of bugs are fixed as soon as detected. The developers and users report about security and performance bugs in bug repositories. Due to huge number of submitted bug reports and lack of tool support for isolating security and performance bug reports, it becomes a time consuming task for bug triager to analyze each bug report and identify these high impact bug reports. As a result, there is chance to overlook these high impact bugs or delay in identifying them. An unnoticed or delay in fixing can cause serious loss (i.e. unauthorized access due to security bug or switching to competitive providers due to performance bug) to software systems. So, automated identification of security and performance bug reports is needed. In this thesis, the machine learning based approach is proposed to automatically identify newly arrived security and performance bug reports. But there is challenge due to class-imbalance phenomenon. The number of security and performance bug reports are smaller compared to other type of bug reports. This thesis proposes textual information based classification to identify security and performance bug reports with feature selection and under-sampling. This thesis also works with structural information of bug reports to identify security and performance bug reports.
Das, Dipok Chandra, and Md Rayhanur Rahman. "Security and performance bug reports identification with class-imbalance sampling and feature selection." 2018 Joint 7th International Conference on Informatics, Electronics & Vision (ICIEV) and 2018 2nd International Conference on Imaging, Vision & Pattern Recognition (icIVPR). IEEE, 2018.