A New Test Case Selection Method Based on Data FlowClustering

摘要

Test case selection is an important activity of regression testing. The purpose is to select a subset from a huge test suite and execute these test cases so as to detect the faults of the modified program. The test case selection process needs to trade-off the scale and fault detection capacity of the test suite. In this paper, we propose a new test case selection method based on data flow clustering. The method first collects program execution information, then the Def-use pairs feature vectors are extracted. Afterwards, we select the test cases from the original test suite by means of the cluster algorithm and sample method. The experimental analysis shows that the method can select the test cases effectively and improve the efficiency of the testing.