Scalable and Accurate Test Case Prioritization in Continuous Integration Contexts

Continuous Integration (CI) requires efficient regression testing to ensure software quality without significantly delaying its CI builds. This warrants the need for techniques reduce time, such as Test Case Prioritization (TCP) that prioritize execution of test cases detect faults early possible. Many recent TCP studies employ various Machine Learning (ML) deal with dynamic and complex nature CI. However, most them use a limited number features training ML models evaluate on subjects which application makes little practical sense, due their small time low failed In this work, we first define, at conceptual level, data model captures sources relations in typical environment. Second, based model, define comprehensive set covers all previously used by related studies. Third, develop methods tools collect defined 25 open-source systems enough builds whose takes least five minutes. Fourth, relying collected dataset containing feature set, answer four research questions concerning collection effectiveness ML-based TCP, impact effectiveness, decay over trade-off between techniques.