AI-Driven Software Testing Optimization: A Comprehensive Machine Learning Framework for Automated Test Generation and Oracle Problem Resolution
International Journal of Innovations in Science and Technology · 02.05.2026 · 258
DOI: 10.33411/ijist/1799 ↗Kurz gefasst
Fachartikel in der Zeitschrift International Journal of Innovations in Science and Technology, online erschienen am 02.05.2026. Verfasst von Zoobia zeeshan und 3 weiteren Autor:innen. Die Arbeit nennt 34 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0).
Abstract (Originalsprache)
Software testing is often one of the most expensive and time-consuming phases in software development, and it becomes even harder to manage in fast and continuous development environments. This study presents an AI-driven software testing framework that combines supervised, unsupervised, and reinforcement learning to improve automated test generation, fault prediction, test prioritization, and test oracle decision-making. The proposed framework is designed not only to improve testing efficiency but also to address the test oracle problem, which involves determining expected outputs for generated test cases. For this purpose, it integrates statistical anomaly detection, metamorphic relation discovery, and behavioral pattern recognition within a single architecture. The model was evaluated on 15 open-source projects and 3 industrial case studies to check its practical usefulness in different testing scenarios. Experimental results show that the framework reduces overall testing time by approximately 34.7% while maintaining a fault detection accuracy of 97.8%. In addition, the proposed oracle mechanism achieved around 94.2% accuracy in identifying test failures, and the reinforcement learning module improved test prioritization effectiveness by approximately 28% compared to baseline methods. These results indicating that hybrid learning approaches can improve adaptability and effectiveness in software quality assurance tasks in one framework can provide a more adaptive and effective solution for modern software quality assurance, although further improvements are required for highly dynamic systems, especially for highly dynamic systems.
Zitieren
APA
zeeshan, Z., A., A., arif, W., & Ain, Q. U. (2026). AI-Driven Software Testing Optimization: A Comprehensive Machine Learning Framework for Automated Test Generation and Oracle Problem Resolution. International Journal of Innovations in Science and Technology, 258. https://doi.org/10.33411/ijist/1799
Vancouver
zeeshan Z, A. A, arif W, Ain QU. AI-Driven Software Testing Optimization: A Comprehensive Machine Learning Framework for Automated Test Generation and Oracle Problem Resolution. International Journal of Innovations in Science and Technology. 2026:258. doi:10.33411/ijist/1799
BibTeX
@article{zeeshan2026,
author = {zeeshan, Zoobia and A., Aamna and arif, wajiha and Ain, Qurat Ul},
title = {AI-Driven Software Testing Optimization: A Comprehensive Machine Learning Framework for Automated Test Generation and Oracle Problem Resolution},
journal = {International Journal of Innovations in Science and Technology},
year = {2026},
pages = {258},
doi = {10.33411/ijist/1799}
}
EUVIMED