Explainable Test Case Prioritization in Continuous Integration through Incremental Learning Approach

In Continuous Integration (CI) environments, where software undergoes frequent updates, regression testing is vital in ensuring software quality. However, rerunning every test case becomes impractical with the constant changes. Test Case Prioritization (TCP) addresses this issue, and Machine Learning (ML) is increasingly used to manage regression testing. However, many ML models struggle to adapt to new features or changes in CI and lack transparency, complicating the regression process. To address these issues, we propose an incremental learning-based explainable ML model for TCP in CI environments, which adaptively incorporates new changes. We use SHapley Additive exPlanations (SHAP) to evaluate feature contributions and help testers understand the model's functionality. Our model is trained and tested on 20 open-source software projects. Its performance is assessed using Accuracy and F1 Score, while test case prioritization is evaluated with the Average Percentage of Faults Detected (APFD) and a new metric, the Failed Test Ranking Score (FTRS).