INTELLIGENT ENSEMBLE LEARNING APPROACH FOR SOFTWARE DEFECT PREDICTION

Abstract

Software defect prediction plays a crucial role in improving software quality by identifying fault-prone modules at early stages of development. With the increasing size and complexity of modern software systems, traditional defect detection techniques often fail to provide accurate and timely predictions. This paper presents an intelligent ensemble-based model for software defect prediction that integrates multiple machine learning classifiers to enhance prediction performance and robustness. The proposed approach combines the strengths of diverse learning algorithms through ensemble techniques such as voting and stacking, enabling more reliable classification of defective and non-defective software modules. Feature selection and data preprocessing methods are applied to handle noise, imbalance, and redundancy in software metrics. Experimental evaluation conducted on benchmark software defect datasets demonstrates that the ensemble-based model significantly outperforms individual base classifiers in terms of accuracy, precision, recall, and F1-score. The results indicate that the proposed intelligent ensemble model effectively reduces false predictions and supports early defect identification, thereby minimizing maintenance costs and improving overall software reliability