Artificial Neural Network (ANN)-Based Classification of Photocatalytic Dye Degradation Utilizing MATLAB

Abstract

This study applied MATLAB-based artificial neural network (ANN) methods to predict photocatalytic dye degradation. Standard experimental techniques need help to capture process complexity, hindering accurate predictions. Leveraging machine learning, specifically ANN, and optimal design and input variables, this research overcame challenges to create a prediction model. Real-life data were used to evaluate the model's accuracy, offering valuable insights into effective and affordable wastewater treatment approaches. The Levenberg-Marquardt (LM) algorithm wasemployed in this study's ANN architecture and was well-suited to non-linear regression problems. Experimental data on photocatalytic dye degradation were used as real-life data that serve as the basis for evaluating the accuracy and performance of the prediction model. The results showed that the ANN models performed well, with R² values ranging from 0.9917 to 0.99814 in training, validation, and testing for the selected structure (3-10-1) predictions for photocatalytic dye degradation. This study successfully developed a highly accurate ANN model for predictingphotocatalytic dye degradation during wastewater treatment, demonstrating machine learning’s promising potential for optimizing such environmental processes.