Optimizing regression test suite reduction

Abstract

Software reliability is one of the essential features in the system quality. The purpose of the regression testing is to establish quality of the modified software before the software deployment. The main issue in this regression model is how to select a subset of test cases from the large test suite in the modified software version. Conventional regression subset selection models have addressed this issue on a small set of test suites or limited test cases. In this paper, we have studied the background work of different hybrid regression models and its limitations on the standard test suite. For different software applications, the proposed methodology is applied to find various combinations of faults randomly injected into various combinations of code lines as specified. We use selenium builder to test the software applications and run it in junit eclipse IDE to find faults. The tested output is interpreted with java randomly induced faults, and the system of methodology is evaluated in terms of APFD and APFD box plot.