International Journal of Computer Applications
Severity is defined as the impact of a failure on software. It reflects upon the degree to which the severity affects the functionality and the degree to which a defect can have impact on the development or operation of software. Defect severity assessment is essential in order to allocate testing resources and effectively plan testing activities. In this paper, the authors use text classification techniques to predict and assess the severity of defects. The results are based on defect description of issue requirements obtained from NASA project. They have used Support Vector Machine (SVM) technique to predict defect severity from issue reports.