Contribution of Temporal Sequence Activities To Predict Bug Fixing Time

Abstract

The bug-fixing process challenges development teams and practitioners for best practices that may pave the way not only to efficient human resources management but also to provide information in advance on the required time to investigate and fix a bug. In this study, we proposed a temporal sequence activity model based on Hidden Markov Models to predict bug fixing time. Comprehensive evaluation results of two different scenarios based on bug reports existing in the the Bugzilla repository were provided. Our experiments demonstrate the feasibility of the proposed model in which the most accurate configuration was obtained with the 50 percent of bug records for training and test set.