A Preliminary Framework for Optimising Test Case Selection Using Natural Language Processing and Test Case Prioritisation Using Deep Learning in Continuous Integration

Abstract

Regression testing, though crucial for software quality, is extremely time consuming especially in continuous integration environments. Test case prioritisation can however improve the efficiency of this process by re-ordering test execution. However, due to the growing volume of test cases in continuous integration, effective selection and prioritisation of test cases is a noteworthy challenge. This paper focusses on improving test case selection and test case prioritisation for regression testing in dynamic continuous integration environments. By applying natural language processing to semantically analyse test cases to include implicit test cases in testing, and deep learning to prioritise the most impactful test cases, this approach will reduce the time and resources required for regression testing while maintaining high fault detection and feedback transmission rates. This early stage research paper proposes a new conceptual framework for optimising regression testing in continuous integration environments using natural language processing and deep learning and outlines potential challenges and future research. The expected outcomes of this proposed framework include improved fault detection rates, reduced overall testing time, and improved resource utilisation. The effectiveness of the framework will be evaluated using metrics such as fault detection rate, testing cycle duration, and computational resource efficiency, providing a comprehensive assessment of its impact on software quality assurance.