Significance of Support Vector Machine Classifier for Predicting Defects in Software Development

Abstract

Software quality is a critical aspect of software development, as defects can significantly affect functionality, security, and user satisfaction. Predicting software defects early in the development cycle enables proactive measures to enhance software reliability and reduce costs. This study investigates software defect prediction using machine learning algorithms, particularly Support Vector Machine (SVM). By leveraging the JM1 dataset from the NASA Promise repository, we explore feature selection techniques, classification models, and performance evaluation metrics as the methodology in this study. The findings of the study revealed that SVM played crucial role in defect prediction in software development with 99% in both accuracy, precision, recall and F1score with optimization, 99.88% accuracy, 99.43% precision, 100% recall and 99.71% of F1 score without optimization. This is invariably  enhance software quality assurance practices.