A predictive analytics approach for software defect prediction using hybrid ensemble learners

Abstract

Detecting defects in software systems is essential for maintaining project stability and efficiency. Identifying issues at an early stage helps prevent failures, improve performance, and optimize the allocation of resources. However, predicting software defects remains a challenge due to the complexity of large datasets and the imbalance between defective and non-defective instances. To address these challenges, a novel machine learning-based framework is introduced that enhances accuracy in defect prediction through an advanced stacking ensemble approach. newlineThis study provides a structured approach to improving software defect detection by integrating multiple classifiers in a hierarchical ensemble, refining data quality, and optimizing hyperparameters. Unlike conventional models that rely on a single classification technique, this method enhances predictive reliability through a comprehensive stacking strategy. The findings highlight the role of ensemble learning and preprocessing in achieving more accurate defect detection, reducing maintenance efforts, and improving overall software quality. newlineFuture research will explore advanced methods such as deep learning, transfer learning, and automated feature engineering to further enhance prediction accuracy and generalization across diverse software environments. This work highlights the importance of continuous innovation in predictive modeling for defect detection, paving the way for more reliable and efficient software development practices. newline

Description

Keywords

Citation

item.page.endorsement

item.page.review

item.page.supplemented

item.page.referenced