ENHANCING BUG TRIAGE ACCURACY IN LARGE-SCALE SOFTWARE PROJECTS VIA DEVELOPER ACTIVITY AND ISSUE REPORT ANALYSIS

Abstract

Bug triage is a critical activity in large software projects because the timely and accurate assignment of issue reports to suitable developers directly affects software maintenance efficiency, defect resolution time, and overall project quality. However, manual bug triage becomes increasingly difficult as projects grow in size, involve distributed development teams, and generate large volumes of issue reports. This paper investigates an intelligent bug triage approach that combines developer activity data and issue report analysis to improve assignment accuracy in large-scale software repositories. The proposed method analyzes textual information from issue reports, including titles, descriptions, severity labels, and component details, together with developer-related features such as prior bug-fixing history, recent commits, expertise areas, response frequency, and workload distribution. By integrating these two sources of information, the model identifies the most suitable developer for each incoming bug report. The results indicate that the combined feature-based approach performs better than traditional issue-text-only methods, achieving higher triage accuracy, improved top-k recommendation performance, and reduced assignment errors. The findings suggest that developer activity patterns provide valuable contextual signals for bug assignment, especially in projects with frequent issue inflow and multiple active contributors. Overall, this study demonstrates that combining issue report analysis with developer activity profiling can support more reliable, scalable, and efficient bug triage in large software projects.