Software Defect Prediction Using an Intelligent Ensemble -Based Model

Software defect prediction is an essential part of software quality assurance that seeks to identify potential issues before they become costly ones. This paper presents a prediction method that uses an intelligent ensemble-based machine learning model to determine if software modules are broken or not. The model uses static code metrics such as Lines of Code, Cyclomatic Complexity, Coupling, and Inheritance Depth to produce predictions. Users can manually enter measurements or upload datasets using the system's flexible and user-friendly interface, which is integrated within a Flask-based web application. To help developers and testers prioritize their efforts, clear predictions and helpful comments are provided. The ensemble model increases reliability while enhancing accuracy and robustness by combining the benefits of many classifiers. This application demonstrates the usefulness of AI in software engineering and serves as a foundation for future developments in automated defect analysis. Keywords: Software Defect Prediction, Ensemble Learning, Machine Learning, Static Code Metrics, Software Quality Assurance