Implementation of Support Vector Machine Algorithm with Particle Swarm Optimization Parameter Tuning for Water Quality Classification

This study investigates the use of Support Vector Machine (SVM) enhanced with Particle Swarm Optimization (PSO) for water quality classification. Conventional SVM models often underperform when parameters are selected manually, resulting in reduced predictive accuracy. To overcome this limitation, PSO was applied to automatically optimize the SVM kernel parameters, enabling more reliable and robust classification. The research employed a quantitative experimental framework consisting of data preprocessing, model training, optimization, and performance evaluation. The dataset included physical and chemical attributes of water quality, which were normalized and prepared before classification. Evaluation was based on standard metrics such as accuracy, precision, recall, and F1-score. The results show that the PSO-optimized SVM consistently outperformed the baseline SVM model, producing more accurate and stable classifications. This confirms the potential of metaheuristic optimization in strengthening machine learning approaches for environmental data analysis. The main contribution of this study lies in applying a PSO–SVM framework to water quality classification, a domain where such integration has been rarely explored despite its importance for sustainable resource management. The findings provide both theoretical implications for advancing metaheuristic applications in environmental informatics and practical benefits for improving decision support in water quality monitoring and management.