Estimation of Prioritization of Test Cases Using Machine Learning Algorithms

Regression testing is required to ensure that modifications to computer code do not disrupt its meaning. Executing all tests may be momentary and resource-intensive with the widespread adoption of Agile Development (CI) in web applications, which increases the frequency of running software builds. To solve this gap, Test Case Selection and Prioritization (TCP) strategies for improving testing activities by selecting and prioritizing test cases to supply real-time feedback to programmers have been developed. Scientists have recently relied on Algorithm (ML) approaches to build successful TCP (ML-based TCP). These strategies assist in combining knowledge about new tests from incomplete and imperfect resources to produce reliable projections. This paper undertakes a systematic literature review of ML-based TCP approaches, to perform an in-depth study of the state of the art and provide insights into future research directions. To that goal, we examine twenty-nine primary papers published between 2006 and 2020 and identify them using a methodical and recorded approach. This work covers five research concerns, including variances in ML-based TCP approaches and global features for training and testing ML models, alternate metrics for assessing techniques, method performance, and repeatability of published findings.