Enhancing Software Engineering Education through AI: An Empirical Study of Tree-Based Machine Learning for Defect Prediction
Ensaf Alhazeem, Anas Alsobeh, Bilal Al‐Ahmad
Abstract
In the rapidly evolving field of information technology education,integrating artificial intelligence (AI) and machine learning (ML) techniques presents opportunities and challenges. This empirical study investigates the application of tree-based ML techniques, specifically Random Forest (RF) and Extreme Gradient Boosting (XGBoost), for software defect prediction in the context of IT education. We analyze nine publicly available NASA software defect datasets to compare the performance of these algorithms across multiple metrics, including accuracy, precision, recall, and ROC area. Our findings demonstrate that XGBoost consistently outperforms Random Forest, achieving near-perfect accuracy across most datasets. The paper explores how these advanced techniques can be responsibly integrated into software engineering (SE) education to enhance student learning while addressing concerns about potential over-reliance on AI tools. We discuss the implications of our results for IT education, emphasizing the need to balance the use of sophisticated AI technologies with the development of fundamental software assurance skills. Furthermore, we examine the role of AI in augmenting SE education, particularly in areas such as software assurance explanations, feature identification, and data augmentation.