A novel method of mutation clustering based on domain analysis

Mutation testing is an effective but operational expensive technique. There exists much improvement to help mutation testing become a wide-used testing technique. The main objective is to reduce the number of mutants and test cases to be executed. The simplified mutant set can still approximate the adequacy. Many new techniques are put forward continuously. One of them is called mutation clustering. It integrates data clustering and mutation analysis, such that both mutant set and test set can be reduced dramatically. This paper revisions the extant mutation clustering. The goal is to provide a method which can guide clustering mutants statically before test case generation. And then the simplified mutant set is used to generate a test set. We hypothesize that such a test set can kill the original mutant set approximately. The cost of mutation testing will further decreases, because only a small test set needs to be generated and executed. A case study shows the encouraging result to support our hypothesis