Effective Bug Assortment Using Data Reduction Techniques

Software companies spend over 48% of their total cost to fix the bugs. An effective way to automatically fix the bugs to the correct developer is called Bug Triage or Bug Assortment. Data sets containing the bug reports are collected from two large open source projects like Mozilla and Firefox. These projects consist of open source bug repositories. Bug repositories are large repositories which stores all the details of bugs. The details are stored in the form of a bug report. These bug report are saved as a document and a related developer is mapped to the label of the document. Software companies spend most of their total cost in fixing these bugs. In bug repositories the two main challenges faced is the large quantity of the data set and the low quality. Noise and redundancy are the main cause for the low quality of the data set. However, irrespective of all these difficulties assigning a proper developer to fix the bug is not an easy task without knowing the actual class of the bug. In this paper we propose data reduction technique which reduces the high scale of the data but it retains the quality of the data set. We also propose domain wise bug solution.