关键词: 测试用例排序, 聚类, 代码单元覆盖矩阵

Abstract: Regression testing is an expensive but important process in software development. As an effective technique in improving testing efficiency, test case prioritization is one of the hot research topics recently. Test case prioritization orders the test cases according to the testing function so that the most beneficial ones are executed first. At the same time, data correlations and patterns in software codes could help improve test case prioritization. Different code units have different testing priorities. Thus, clustering algorithms can help to analyze the structure information in test cases and induct the test case prioritization. Experimental results show that, the improved scheduling strategy can improve the performance of clustering based test case prioritization.