Severity Prediction of Bug Reports Using Weighted Implicit Tags

Abstract

Bug reports are crucial in the software development and maintenance of large-scale software projects. Although several bug reports may be received daily, the severity of bug reports influences the fixing priority. In this paper, we consider the abundant information embedded in tags, and propose a tagging approach for assigning weighted implicit tags to bug reports by calculating the report similarity. The experimental results show that the proposed tagging approach can effectively improve the prediction performance of bug report data sets with comprehensive tag information. For the bug report data sets with poor tag information, the proposed tagging approach can improve the performance in most cases.