Review on Leveraging Techniques on Bug Repository to form Accurate Bug Triage [ ]


Software organizations spend huge amount of cost on managing programming bugs. An unavoidable stride of fixing bugs is bug triage, which expects to effectively allocate a developer to a new bug. To diminish the time cost in manual work, text classification techniques are applied to perform automatic bug triage. In this paper, we address the problem of information decrease for bug triage, i.e., how to diminish the scale and enhance the nature of bug data. We use instance selection with feature selection at the same time to decrease information scale on the bug dimension and the word dimension. To focus the request of applying instance selection and feature selection, we extract properties from historical bug information sets and construct a predictive model for new bug information set. Outcomes demonstrate that our data reduction can adequately decrease the data scale and enhance the precision of bug triage. Our work gives a way to deal with leveraging techniques on data processing to form decreased and high-quality bug information in programming advancement and upkeep.