Enhancing photocatalytic degradation of hazardous pollutants with green-synthesized catalysts: A machine learning approach.
Journal:
Journal of environmental management
Published Date:
Jun 1, 2025
Abstract
The effective removal of organic pollutants from wastewater necessitates the development of advanced photocatalytic materials. This study explores the application of machine learning algorithms to predict the degradation efficiency of PRM using green-synthesized photocatalysts. Utilizing a variety of machine learning models-including Ridge Regression, Random Forest, Gradient Boosting, XGBoost, and Bagging Regressor- among others, the analysis identified models capable of achieving high accuracy and reliability in predicting PRM degradation under varying conditions. Among the models tested, Gradient Boosting and XGBoost with GridSearchCV optimization demonstrated the highest performance, achieving R values exceeding 0.99 along with significantly lower MSE and MAE, confirming their ability to capture complex degradation patterns with high fidelity. While linear models such as Ridge regression performed reasonably well, they exhibited limited flexibility compared to nonlinear ensemble methods. The synthesis of ZnO and Ni-doped ZnO photocatalysts using eco-friendly techniques underscores the importance of sustainable approaches in environmental applications. The integration of machine learning with green chemistry principles not only advances wastewater treatment technologies but also highlights the potential of interdisciplinary approaches in developing sustainable environmental solutions. This work demonstrates that combining eco-friendly synthesis with advanced ML models can significantly enhance predictive accuracy, guiding the optimization of photocatalytic processes for practical environmental remediation.
