Enhancing supervised bug localization with metadata and stack-trace

• Published in: Knowledge and information systems Vol. 62; no. 6; pp. 2461 - 2484
• Main Authors: Wang, Yaojing, Yao, Yuan, Tong, Hanghang, Huo, Xuan, Li, Ming, Xu, Feng, Lu, Jian
• Format: Journal Article
• Language: English
• Published: London Springer London 01-06-2020; Springer Nature B.V
• Subjects: Accuracy; Analogies; Computer Science; Data Mining and Knowledge Discovery; Database Management; Debugging; Information retrieval; Information Storage and Retrieval; Information Systems and Communication Service; Information Systems Applications (incl.Internet); IT in Business; Localization; Metadata; Modelling; Regular Paper; Software development; Bug reports; Supervised topic modeling; Metadata; Bug localization; Stack-trace
• ISSN: 0219-1377; 0219-3116
• DOI: 10.1007/s10115-019-01426-2

Locating relevant source files for a given bug report is an important task in software development and maintenance. To make the locating process easier, information retrieval methods have been widely used to compute the content similarities between bug reports and source files. In addition to content similarities, various other sources of information such as the metadata and the stack-trace in the bug report can be used to enhance the localization accuracy. In this paper, we propose a supervised topic modeling approach for automatically locating the relevant source files of a bug report. In our approach, we take into account the following five key observations. First, supervised modeling can effectively make use of the existing fixing histories. Second, certain words in bug reports tend to appear multiple times in their relevant source files. Third, longer source files tend to have more bugs. Fourth, metainformation brings additional guidance on the search space. Fifth, buggy source files could be already contained in the stack-trace. By integrating the above five observations, we experimentally show that the proposed method can achieve up to 67.1% improvement in terms of prediction accuracy over its best competitors and scales linearly with the size of the data.