Machine Learning-Based Adaptive Test Sequence Recommendation System for Regression Testing

Abstract

In software testing, especially regression testing, selecting test sequences manually can take a lot of time and may lead to mistakes, making the process less effective. This study introduces the Adaptive Test Sequence Recommendation System (ATSRS), which uses machine learning to suggest the best test sequences. The system analyzes how different parts of the software interact and uses this information to make its recommendations. It continuously learns and improves by fine-tuning its data using linear. The ATSRS is trained with this improved data using a boosting method that combines several simple decision tree models into a stronger one. The experimental study has shown that the system is very accurate, with over 98% precision in recommending the right test sequences. This high level of accuracy makes regression testing more effective and efficient. By providing relevant and effective test sequences, the ATSRS helps improve the overall process of software testing.