An Efficient Software Fault Prediction Model using Spider Wasp Optimization and Ensemble Learning

Abstract

By spotting flaws early in the development process, Software Fault Prediction (SFP) is crucial for increasing software dependability. In order to address the drawbacks of conventional techniques like the Genetic Algorithm and Particle Swarm Optimization, this research suggests an improved feature selection strategy utilizing the Spider Wasp Optimization (SWO) algorithm. To increase prediction accuracy, machine learning models such as Decision Tree, Naïve Bayes, KNN, LDA, and XGBoost are used to assess the chosen best features. Results from experiments on the KC1 dataset show that the SWO-based method outperforms all classifiers. Additionally, a web interface built on Flask is created to provide real-time software defect prediction.