Bug Triage System with Bug Data Analysis

Software bugs are the major areas where most of the IT companies spend its 45% of efforts. The most efficient way of reducing bugs in the software is bug triage, which is appropriately assigning the new but to a developer. Automatic Bug triage is conducted by using text classification methods to minimize the manual work. In this paper, the problem of huge data for bug triage is addressed i.e. reducing the amount of data in assigning the bugs to developers. This technique combines feature selection and instance selection for simultaneously minimizing the bug dimension and word dimension. To identify the application order of instance selection and then feature selection, extraction of various attributes from previous bug track record is carried out to build new bug track or dataset. The proposed system tests its performance on the sample dataset of the open source software i.e. Eclipse. The proposed system works on novel data processing techniques to generate small amount of but highly informative bug report data in software maintenance and development.