Review on Domain Specific Bug Triage with Software Data Reduction Techniques

Abstract: Bugs are very essential aspects in a software company. The process of fixing bugs is called as a bug triage. Bug triage is an unavoidable step in a software company. In bug triage a correct developer is given to a new bug for fixing it. To manually perform the bug triage is very costly and even time consuming. So text classification techniques are used which uses automatic bug triage. There is a problem of large data i.e the data should be reduced and the quality of the data should be increased. To perform this instance selection and feature selection are used simultaneously. For this we should know the order for applying instance selection and feature selection, and to know the order we extract the attributes from the bug data sets. For the experiments we are using two open source projects such as eclipse and Mozilla. And our result shows that the data is reduced with high quality bug data sets.

Keywords: Mining Bug repositories, bug data reduction, attribute extraction, instance and feature selection.