SOFTWARE DEFECT PREDICTION USING INTELLIGENT ENSEMBLE BASED MODEL

Abstract

Software defect prediction is a critical task in software engineering, aiming to identify potential defects in software modules before deployment. Accurate defect prediction helps in improving software quality, reducing maintenance costs, and optimizing testing efforts. Traditional machine learning models have been widely used for this purpose, but their performance often varies across datasets due to data imbalance and feature complexity. To address these challenges, this study proposes an intelligent ensemble-based model for software defect prediction. The ensemble approach integrates multiple base learners, combining their strengths to enhance prediction accuracy and generalization capabilities. The proposed model incorporates feature selection techniques to eliminate redundant and irrelevant features, thereby improving learning efficiency. Various machine learning algorithms, including decision trees, support vector machines, and deep learning models, are combined using an optimized weighting strategy to maximize predictive performance. The experimental evaluation is conducted on publicly available defect datasets, and results demonstrate that the ensemblebased model outperforms individual classifiers in terms of precision, recall, and overall classification accuracy. This approach provides a robust and scalable solution for defect prediction, aiding software developers in making informed decisions. Future research will focus on further optimizing ensemble strategies and exploring deep learning advancements for enhanced defect prediction.