Software Fault Prediction Using Data Reduction Techniques

Jagruti R. Patil, Swapnaja Suryawanshi

Abstract: The procedure of altering bug will be bug triage, which plans to effectively relegate an engineer to another bug. Programming organizations spend the greater part of their expense in managing these bugs. To decrease time and cost of bug triaging, we display a programmed way to deal with foresee a designer with significant experience to comprehend the new coming report. In proposed approach we are doing information decrease on bug information set which will lessen the size of the information and also build the nature of the information. We are utilizing example determination and highlight choice at the same time with verifiable bug information. We have included another module here which will depict the status of the bug like whether it doled out to any engineer or not and it is redressed or not. The objective of bug triaging is to allot possibly experienced designers to new-coming bug reports. To diminish time and cost of bug triaging, we introduce a programmed way to deal with anticipate an engineer with important experience to understand the new coming report. In this paper, we examine the utilization of five term determination techniques on the precision of bug task. Moreover, we re-adjust the heap between engineers taking into account their experience. We lead investigates four genuine datasets. The exploratory results demonstrate that by selecting a little number of separating terms, the F-score can be altogether made strides.

Keywords: Bug, Bug triage, Data Reduction, Instance selection, Data Mining