Copyright © 2026 Authors retain the copyright of this article. This article is an open access article distributed under the Creative Commons Attribution License which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.

@article{195043,
        author = {G RAGHU VAMSHI and K M KABILAN and PRAHARSH SAI and T. LAVANYA and Dr. S. SHIVA PRASAD},
        title = {A NOVEL APPROACH TO IMPROVE SOFTWARE DEFECT PREDICTION ACCURACY USING MACHINE LEARNING},
        journal = {International Journal of Innovative Research in Technology},
        year = {2026},
        volume = {12},
        number = {10},
        pages = {5996-6008},
        issn = {2349-6002},
        url = {https://ijirt.org/article?manuscript=195043},
        abstract = {Defect prediction is an important area of research in software engineering. Identifying defects early can significantly improve software quality and lower development costs. The goal of software defect prediction is to spot potential faults in source code before the testing phase. This is done by using data mining and machine learning techniques. This research connects software engineering and data analysis to improve defect prediction performance, focusing on feature selection and effective machine learning models. Five publicly available NASA datasets, CM1, JM1, KC2, KC1, and PC1, are analyzed using various classifiers, including Random Forest, Logistic Regression, Support Vector Machine, Decision Tree, and ensemble voting methods. Feature selection is applied to optimize the datasets and increase predictive accuracy.  Besides basic defect classification, the proposed system offers useful predictive insights like defect risk scores, severity levels, and criticality status. This enables more informed and actionable decision-making. The WEKA machine-learning workbench is used for data preprocessing, model training, and evaluation. Minitab is used for statistical validation. Experimental results show a significant improvement in defect prediction accuracy when feature selection is used. The ensemble voting classifier performs the best among all the models tested. Additionally, the system features a modern user interface that improves visualization, clarity, and interpretability of results, making the approach practical and effective for assessing software quality in the real world.},
        keywords = {Software Defect Prediction, Feature Selection, Machine Learning, NASA Software Metrics Datasets, Ensemble Voting Classifier, Defect Risk Assessment, Severity Analysis, Predictive Analytics, WEKA, Software Quality Assurance, Modern User Interface},
        month = {March},
        }