Artificial intelligence (AI) has gathered significant attention and acceptance across various domains, including water treatment and desalination, where it has proven to be a valuable tool for enhancing process efficiency and addressing water pollution and scarcity. AI techniques offer optimized chemical usage, reduced operational costs, and effective solutions. In this paper, a hybrid method, SVR-PSO is proposed for modeling the dye removal process. First, Particle Swarm Optimization (PSO) tunes the best hyperparameters for Support Vector Regression (SVR). Subsequently, Support Vector Regression (SVR) is built using the obtained best hyperparameters and then fitted. The proposed method is tested and evaluated for modeling the removal of The Malachite Green dye using Helianthus Annuus seed shells as an eco-friendly adsorbent. The proposed method is compared with SVR tuned with Genetic Algorithm (SVR-GA) and SVR tuned with the Grid Search method (SVR-GS), and the obtained results demonstrate the efficiency of the proposed method.

Support Vector Regression, Particle Swarm Optimization, Genetic Algorithms, dye removal.