Automatic Fine-Grained Issue Report Reclassification

摘要: Issue tracking systems are valuable resources during software maintenance activities. These contain different categories of issue reports such as bug, request for improvement (RFE), documentation, refactoring, task etc. While logging into a system, reporters can indicate the category reports. Herzig et al. Recently reported that more than 40% given wrong in systems. Among marked bugs, 30% them not bug The misclassification adversely affects developers they then need to manually identify various To address this problem, paper we propose an automated technique reclassifies report appropriate category. Our approach extracts feature values from and predicts if needs be reclassified its We have evaluated our reclassify 7,000 HTTP Client, Jackrabbit, Lucene-Java, Rhino, Tomcat 5 1 out 13 categories. experiments show achieve weighted precision, recall, F1 (F-measure) score ranges 0.58-0.71, 0.61-0.72, 0.57-0.71 respectively. In terms F1, which is harmonic mean precision substantially outperform several baselines by 28.88%-416.66%.