| Detection and correction of software faults is a difficult and complex issue. Mining software repositories is one of the approaches to assess and improve software quality.; The objective of our research was to mine software quality data from a large-scale open-source software system, Mozilla and its associated defect tracking database, Bugzilla.; The first contribution in this dissertation is the development of Entity Relationship diagram of Bugzilla database by performing data reverse engineering.; The second and most important contribution in this research is the development of Metrics-Delta data set, based on Mozilla source code and Bugzilla database. The data set contains 22 attributes and 8349 data points.; The final contribution is development of predictive models using support vector machine, neural networks and linear regression for Metrics-Delta data set. Predictive models can be used to predict software quality, which in turn minimizes failure rate and could be used as a management tool. |
