Date of Graduation
Statler College of Engineering and Mineral Resources
Lane Department of Computer Science and Electrical Engineering
In practical software development projects, solving test issues efficiently during Software Development Life Cycle is critical to release software products on time. Different test environments, test resources and test requirements could result in different outcomes. Therefore, getting accurate prediction of the software defects' resolution time could be beneficial to the practical projects.;In our study, data mining techniques offer great promise in prediction of software defects' resolution time. Our research is conducted based on the NASA Metrics Data Program (MDP). We first calculate the resolution time for available projects. Using unsupervised discretization methods, we split resolution time into certain interval as response variable. Then, investigating the relationship between metric properties and time intervals, we fit a model that attempts to produce prediction on resolution time. Experiments and analysis successfully demonstrate the feasibility of our approach.
Wang, Da, "Prediction for Resolution Time of Software Defect" (2010). Graduate Theses, Dissertations, and Problem Reports. 3072.