Machine Learning for Test Case Prioritization in Continuous Integration: A Comprehensive Analysis

This study introduces an innovative Predictive Test Prioritization (PTP) methodology for Continuous Integration (CI), utilizing historical test case execution data. To predict the probability of success or failure in new test cases, machine learning classifiers like k-nearest Neighbors, Random Forest, Support Vector Machine (SVM), Gradient Boosting, and Logistic Regression are applied and trained on historical data. The evaluation process encompasses metrics such as F1-score, recall, accuracy, and precision, offering a nuanced understanding of the effectiveness of the classifiers. The overarching goal is to optimize test prioritization, potentially enhancing software testing efficiency. This research offers valuable insights into continuous integration systems, emphasizing the pivotal role of predictive strategies in refining testing practices and contributing to the knowledge base in CI.