AI-Based Test Case Prioritization Using Defect Risk Prediction

Software testing often operates under time constraints that prevent the exhaustive execution of all test cases. Test case prioritization aims to address this challenge by ordering test cases to detect defects as early as possible during execution. This paper proposes an AI-based test case prioritization approach that leverages machine learning-driven defect risk prediction. Defect probabilities are first estimated at the module level using static code metrics and a supervised learning model. These risk scores are then propagated to test cases based on their associated module coverage to determine execution priority. The proposed approach is evaluated using a simulated test–module mapping and compared against baseline execution strategies. Experimental results demonstrate that the AI-based prioritization approach achieves higher fault detection effectiveness, as measured by APFD, indicating its potential to support risk-based test automation.