Risk-Based Test Case Prioritization by Correlating System Methods and Their Associated Risks

Regression testing aims to ensure the quality of a software after modification. However, re-executing the entire test suite in regression testing is a time-consuming, costly, and tedious process that often requires additional budget and time. Thus, limited resources always result in early termination and poor quality software. Test case prioritization aims to improve regression testing by re-scheduling the test cases in a manner that could increase the fault detection rate. Risk-based testing has gained popularity in the area of software testing. However, most of the existing methods compute risk values manually, which makes these methods tiresome, laborious, and slow. In this paper, we propose a semi-automatic risk-based test case prioritization approach based on software modification information and methods (functions) invocation relationship. The objective of this research is to make risk-based testing more systematic and flexible by automating the risk assessment process, and find high-risk faults early. The proposed approach utilizes the requirements modification information, complexity, and size of the methods as the risk indicating factors. We applied an automated procedure to extract these risk factors and compute the risk values of the system methods. The proposed approach is empirically evaluated with two software applications with multiple versions in terms of its fault detection rate, both overall and in risky modules. The experimental results suggest that our proposed approach improves test efficiency by spotting defects early overall and even earlier in the high-risk modules than existing state-of-the-art approaches.