基于关联规则挖掘的回归测试选择策略研究-计算机应用技术专业论文.docxVIP

AbstractSoftware Abstract Software testing is one of the most important parts in the process of software development．Regression testing is a kind of test for the purpose to validate modified software correctness，thus the process of running test case in the regression testing phase is the same as that of running test case in other testing phase，only the test suite be different． Recently,regression testing strategy can be divided into“retest-all”and “selective”．With the accretion of software scale，“retest—all”strategy is not feasible both in consideration of the time and cost，especially when there is a minor change． As changed parts may affect other unchanged parts，proper test case selection strategy must be used when regression testing are carried out．How to choose a test suite as small as possible that can cover all the changed parts and unchanged parts affected by changes is one of the most popular research field．Developers can tell testers that which parts were changed when a new version of software appears．It simplifies the process of choosing test cases which can cover all the changed parts．However,it is much more difficult to choose all the test cases which belong to unchanged paas actually affected by changes． We assume that test cases cover either changed parts or unchanged parts affected by changes．In order to construct a much more accurate regression test suite and improve the efficiency of regression test,we propose a new test case selection strategy． The strategy is as follows．Firstly,select all test cases that cover the changed parts． Secondly,by applying association rules mining to the remaining ones in baseline test cases，we try to find some rules among test cases，and according to these rules we can exclude some test cases from regression test suite．Finally,we prioritize test CaseS according to the number of modules and line of code they coveL By experiment we can conclude that this method not only chooses all test cases that cover the changed parts，bu