Classifying software issue reports through association mining

Software issue reports classification is a significant task in software maintenance and evolution. Despite the research effort being made over the years, the existing issue reports classification techniques are still inadequate. In this paper, we propose a new approach that is inspired by the Classification Associations Rule Mining (CARM) methodology in data mining, and report the testing of our method on 500 software issue reports extracted from an open source issue tracking system. Our experiments show that our method can achieve a high degree of accuracy in classifying software issue reports