SOFTWARE DEFECT PREDICTION USING AN INTELLIGENT ENSEMBLE - BASED MODEL

Abstract

Software defect prediction is a critical task in software engineering aimed at identifying faulty modules before deployment to improve software quality and reduce maintenance costs. Traditional defect detection methods rely heavily on manual testing and static analysis, which are time-consuming and often fail to detect complex defects. This project proposes a Software Defect Prediction System using an Intelligent Ensemble-Based Model, which leverages multiple machine learning algorithms to enhance prediction accuracy and reliability. The proposed system utilizes historical software metrics such as code complexity, lines of code, coupling, cohesion, and change history to predict defect-prone modules. Individual models such as Decision Tree, Random Forest, Support Vector Machine (SVM), and Neural Networks are trained on the dataset. These models are then combined using ensemble techniques such as bagging, boosting, or voting to create an intelligent hybrid model. The ensemble approach improves performance by reducing overfitting and capturing diverse patterns in the data. The system is evaluated using performance metrics such as accuracy, precision, recall, F1-score, and ROC-AUC. Experimental results demonstrate that the ensemble-based model outperforms individual classifiers in predicting software defects with higher accuracy and lower error rates. This approach enables early detection of defects, reduces development costs, and improves software reliability. Overall, the project highlights the effectiveness of ensemble learning in building robust and efficient software quality prediction systems.