A Survey on Software Data Reduction Techniques For Bug Triage Process

ABSTRACT Software companies are trade with large number of software bugs. It is well exclusive and mandatory too. Bug triaging process is nothing but to assign effective and proper developer for bug fixing. In software development, the new bugs are manually triaged by expert developer i.e. human triager. Due to the large number of daily bugs and the lack of know-how, the manual bug triage is expensive in time cost and low in correctness. To avoid the expensive cost in manual bug triage, an automatic bug triage approach is used to predict developers for bug report. Data reduction technique is used to build a small scale and high quality set of bug data for bug triage. The arrangement of feature selection and instance selection techniques are used to reduce the word dimension and bug dimension of data scale. Attributes are extracted from historical bug data sets and a predictive model for a new bug data set is build to find the order of applying feature selection and instance selection.