arXiv.itsnick.co

an independent field guide to 23 million papers from arXiv & OpenAlex — not affiliated with arXiv.org

Social Sciences · Sociology and Political Science

Automated, Highly-accurate Bug Triaging Using Machine Learning

Pamela Bhattacharyaa, Iulian Neamtiua, Christian R. Sheltona

See on the map

Abstract

Empirical studies indicate that automating the bug assignment process (also known as bug triaging) has the potential to significantly reduce software evo-lution effort and costs. Prior work has used machine learning techniques to automate bug triaging but has employed a narrow band of tools which can be ineffective in large, long-lived software projects. To redress this situation, in this paper we employ a comprehensive set of machine learning tools and analyzes that lead to very accurate predictions, and lay the foundation for the next gen-eration of machine learning-based bug triaging. Our work is the first to exam-ine the impact of multiple machine learning dimensions (classifiers, attributes, and training history) on prediction accuracy in bug triaging. We employ four classifiers and perform an ablative analysis to show the relative importance of classifiers and various software process attributes on bug triaging accuracy. We propose optimization techniques that achieve high prediction accuracy while reducing training and prediction time. We validate our approach on Mozilla and Eclipse, covering 856,259 bug reports and 21 cumulative years of develop-ment. We demonstrate that our techniques can achieve up to 86.09 % prediction accuracy in bug triaging and significantly reduce tossing path lengths.

More like this