A Comparison of Summarization Methods for Duplicate Software Bug Reports

초록

Bug reports vary in length, while some bug reports are lengthy, others are too brief to describe bugs in detail. In such a case, duplicate bug reports can serve as valuable resources for enriching bug descriptions. However, existing bug summarization methods mainly focused on summarizing a single bug report. In this paper, we focus on summarizing duplicate bug reports. By doing so, we aim to obtain an informative summary of bug reports while reducing redundant sentences in the summary. We apply several text summarization methods to duplicate bug reports. We then compare summarization results generated by different summarization methods and identify the most effective method for summarizing duplicate bug reports. Our comparative experiment reveals that the extractive multi-document method based on TF-IDF is the most effective in the summarization. This method successfully captures the relevant information from duplicate bug reports, resulting in comprehensive summaries. These results contribute to the advancement of bug summarization techniques, especially in summarizing duplicate bug reports. © 2023 by the authors.

키워드