A 2-Layer Ensemble Machine Learning Based System for Software Defect Prediction

Software defects are anomalies in software projects that lead to incorrect outcomes and deviate from predefined requirements set by customers, developers, and other stakeholders. These defects pose significant challenges and can cause substantial losses if left unresolved. Detecting software defects in the early stages of development is crucial, as it can save considerable resources and reduce costs. To address this, we propose a machine learning model capable of automatic defect prediction at the early stages of development. Our approach introduces a two-layer ensemble-based system that leverages static code analysis by extracting software testing metrics. The proposed double-layer ensemble model aims to surpass the predictive accuracy of standalone models and traditional ensemble techniques, ensuring higher reliability in defect prediction. By implementing this advanced methodology on static code, our research demonstrates the potential of achieving improved defect detection performance, thereby contributing to more efficient and cost-effective software development processes.