Utilizing Association Rule Mining algorithms to Identify Dependencies between Software Defects

Abstract

Focusing on identifying hidden patterns and dependencies between software defects that are difficult to detect using traditional analysis methods, this study employs Association Rule Mining (ARM) to analyze over 140,000 open-source GitHub issues. By leveraging ARM, we have been able to extract explicit association rules that illustrate the interrelations among various issue attributes such as labels and release versions. Our findings indicate strong, meaningful associations that equip developers and quality assurance teams with the information necessary to strategically prioritize issues. This prioritization significantly influences bug fixing and system enhancements. Offering a marked improvement over conventional analysis techniques, ARM provides a clearer and more precise understanding of defect dynamics, thereby enhancing the efficiency and effectiveness of software development practices.