A Hybrid Swarm Intelligence Technique for Feature Selection in Support Vector Machine (SVM) Classifier using Particle Swarm Optimization (PSO) and Ant Colony Optimization (ACO)

Complex optimization issues have been successfully resolved by swarm intelligence algorithms like Ant Colony Optimization (ACO) and Particle Swarm Optimization (PSO). A preprocessing technique called feature selection chooses the most important attribute from datasets in order to minimize their dimensionality. In this research, the Support Vector Machine (SVM) classifier is trained using the features selected by both PSO and ACO. These optimization methods are assessed using performance criteria, including F1 score, accuracy and precision. Additionally, the execution time of both algorithms is measured. A hybrid approach combining PSO and ACO with SVM is explored to further improve feature selection and achieve better overall classification results. This hybrid methodology would maximize the strengths of both algorithms, providing a more robust solution for complex optimization problems. The findings demonstrate the effectiveness of swarm intelligence technique for feature selection, offering an efficient method to enhance model performance.