A Dynamic Metrics Method for Test Case Reuse Based on Bayesian Network

Abstract

How to measure reuse capability of a test case is critical in software test reuse research. A dynamic method based on Bayesian network is proposed for measure the reuse of test case. It collects the information of user's searching and reusing test cases, and then analysis the test case's reuse capability which give basis for consulting to users. The prominent characteristic of this method is that the reuse of test case is divided into three aspects in terms of different reuse preference that improve the accuracy of metric. As the number of test cases available increases, it is becoming more important to devise test case metrics to measure its reuse. The method of measuring test case has two kinds that one kind is reuse metrics basis of test case used information and the other kind is reusability metrics basis of test case itself. In this paper, test case reuse metrics is researched. It is obviously that the more test case is reused, the better reuse of the test case is. According to the behavior characteristic of person, the reuse preference is the most in keep with the reused test case which user chose recently. Hence, all of test case may be taken as black-box without understanding the test case content and is measured by information of users. A dynamic method based on Bayesian network for test case library is proposed which correctly and objectively measure the test case reuse in various customizations of user's reuse preference. To support real-time update, this method recorded user's operation information, and then the corresponding Bayesian network is updated.