Abstract: Bug is a very essential factor which is occur in software. Almost every company facing problem to solve issues related bug, software companies spends over 40-45 percent of cost in dealing with software bugs. Very important step to handle bug is bug triage, which aims to correctly assign a developer to a new bug. To lower the time price in guide work, text category strategies are implemented to behavior computerized bug triage. In this paper, we address the problem of facts reduction for bug triage, i.e., a way to reduce the scale and enhance the satisfactory of bug records. We integrate instance selection with feature selection to simultaneously lessen facts scale on the malicious program dimension and the word measurement. To determine the order of making use of example selection and function choice, we extract attributes from historic malicious program information units and build a predictive model for a new computer virus information set.

Keywords: Mining software repositories, application of data preprocessing, data management in bug repositories, bug data reduction, feature selection, instance selection, bug triage, prediction for reduction orders.