A Novel Hybrid Approach to Enhance Software Quality by Using the Honey Badger Algorithm (HBA) and Support Vector Machine (SVM)

Software defect prediction is the process of identifying modules in software that are likely to be prone to defects. It used to improve software quality and its primary goal is to identify defective modules and send them for testing. Traditional methods of software testing are not capable of performing testing effectively. Effective testing plays a crucial role in software quality assurance. In this paper, we introduce a hybrid approach, Honey Badger Algorithm (HBA) with Support Vector Machine (SVM) for predicting software defects. The HBA finds the most relevant data from a large dataset. It explores all possible data combinations, removes redundancy, and reduces dimensionality. SVM is a supervised machine learning algorithm used for classification, regression, and outlier detection. It is widely used in various fields, including software defect prediction (SDP), due to its effectiveness in handling high-dimensional datasets and complex decision boundaries. When we perform experiments on a given dataset using a proposed model of HBA-SVM, it gives effective and better results than other machine learning algorithms SVM, KNN, and NB. It shows a remarkable number in accuracy 95 %, precision 98%, Recall 44%, F-measure 56%, and AUC score 76%. This result shows that the hybrid HBA-SVM model improves software defect prediction performance and reduces computational complexity.