Abstract:To improve software quality, developers often open a bug repository and allow users to find bugs, describe bugs in the form of bug reports and submit bug reports to the repository. Based on the description, testers assign a priority to each bug report. In the beginning the process of priority assignment is performed manually. With the increasing amount of bug reports, researchers introduced classification methods to assign priorities automatically with all the features considered. In this paper feature selection methods are introduced to improve the effect of bug report prioritization using classification models. The experimental results show that feature selection based on Information Gain and Pearson Correlation can improve the precision and recall for bug report prioritization on two models, i.e., SVM and Naive Bayes.

Exploring the Influence of Feature Selection Techniques on Bug Report Prioritization