Automatic Bug Triaging Process: An Enhanced Machine Learning Approach through Large Language Models

Bug resolution and maintenance are the most critical phases of the software development life cycle. The traditional bug triaging concept refers to the manual assignment of bugs to the appropriate developer after reading the bug details from the bug tracker and further resolving it. The advent of machine learning algorithms provides various solutions for automated bug triaging. Machine learning algorithms can be used to classify bugs and assign each to a developer. Reducing manual efforts optimizes bug-triaging by utilizing manpower in other software development processes. Furthermore, machine learning Large Language Models (LLMs) can be used to take advantage of their natural language processing features and capabilities. This study proposes a machine learning-based embed chain LLM approach for automatic bug triaging. This approach is used to automatically classify bug reports. Based on the results, the appropriate developer is recommended. In addition, the proposed approach is used to automatically predict the priority of bug reports. This paper also discusses the strengths and challenges of the proposed approach.