One-Class Classification Based Bug Triage System to Assign a Newly Added Developer

Abstract

Bug triage is a software engineering problem in which a developer is assigned to a bug report. The existing methods use social network analysis, topic modeling, mining repositories, machine learning, and deep learning. Deep learning methods have shown promising results. However, these methods can not assign a newly added developer to the bug report. In this paper, we proposed a one-class SVM based method that trains a separate classifier for each developer. If a new developer is added to the project, then he or she can be considered by the triage system by training a classifier on his history. The results show an acceptable accuracy score. We believe that our preliminary study can pave the way to address this challenging problem.