A Fault Localization Method Based On Combinatorial Design Approach For Testing

With the continuous development of software technology, the scale of software becomes larger and larger, and the difficulty of software quality assurance also increases. During the software development process, software products often hides all types of faults and defects. Testing for software becomes essential but time-consuming job, and after testing, fault localization is the most important and difficult step. When testing the system under test, there often requires a lot of test cases. Most existing fault localization methods only concern the process and result of failure detection, ignore the relevant links between the large number of the test datum. Data mining can digs out hidden useful information among the large amount datum. Then it can apply data mining to fault localization, and find the relevant information for more efficiently detecting failure.This paper studies a fault localization based on combinatorial test. It finds that failure test cases are caused by a combination of a few parameters. To find the implied relationship among the parameters that causes failure, introducing the association rules algorithm of data mining, and studying the apriori algorithm of association rules. Apriori algorithm can find the frequent item sets from a set of datum.Based on the above research, this paper presents a fault localization based on apriori algorithm approach for testing. This method uses the concept of the minimal failure-causing schema(MFS) of fault localization based on combinatorial test, and through digging the relevance of the parameters from failure test cases, finds the minimal failure-causing schema that causes failure. Compared to the previous methods, the workload significantly reduces.Finally, this method is applied to the experiments. It proves the effectiveness of this method by the experiments. This method can locate failure accurately with less workload, so it has a certain practical value.