Failure history data-based test case prioritization for effective regression test
Jeongho Kim, Hohyeon Jeong, Eunseok Lee · 2017
For regression testing in the continuous integration environments, the time and cost should be considered; to satisfy these constraints, it is necessary to improve the test efficiency regarding the achievement of the test goal. It is especially important to identify the problem quickly by first executing a test case with a high probability of failure. This paper therefore proposes the FHD (Failure History Data)-Prioritization technique for the purpose of effective regression testing. This technique uses the failed test case history, the flipped result of method test case and the correlation data as the prioritization criteria, and the algorithm is designed to calculate the weight through the following two-step classification: 1) The FHD-Prioritization analyzes the failure history data statistically and sequentially arranges the test cases from the highest failure occurrence probability in the current session. 2) If a failure occurs during the test, the FHD-Prioritization reprioritizes in real time based on the correlation data. The performance of the FHD-Prioritization technique is evaluated with Tomcat and Camel, Apache open source software projects that were developed in the continuous integration environment. Because all of these projects are composed of the real faults of real-world projects, it is possible to practically evaluate the efficiency of the proposed approach. The FHD-Prioritization improved the efficiency of test case prioritization by about 5.62+ and 2.17+, respectively, compared to ROCKET and AFSAC.