A Software Regression Testing Strategy Based on Bayesian Network

The purpose of regression testing is to assure quality of the changed system not only due to fixing bugs but also software changes by re-execution a suitable test cases set which determines the quality of such activity. Because of the software complexity and other reasons, there are many uncertainties on selecting regression test case set. In this paper, we present a solution by applying Bayesian Network. Appling the Bayesian Network, model the system requirements and test cases, and define the dependent relationship of requirements. Attain the test suite of regression testing optimized by this way. And the Bayesian Network can be changed by learning constantly in order to optimize the test suite again and again.