Modeling of Groundwater Quality Index Using Multilayer Perceptron (MLP) Neural Network in Al-Mahawil District, Iraq

Abstract

Monitoring water quality is essential for environmental protection and sustainable water resource management. This study employed a Multilayer Perceptron (MLP) neural network to model the Groundwater Quality Index (GWQI) based on key physicochemical parameters. The model was developed using MATLAB’s Neural Network Toolbox (nftool), with data divided into training, validation, and testing sets. Performance was evaluated using MSE, RMSE, MAE, and the correlation coefficient (R). The MLP model achieved good predictive accuracy:(MSE = 2.3, R ≈ 1.0) in training;( MSE = 7.6983, R = 0.99992) in validation; and (MSE = 15.8, R = 0.997) in testing. These results confirm the model’s ability to capture nonlinear relationships and generalize well without overfitting. The stability of the model is supported by its error distribution and training convergence. A comparison with previous studies shows improved performance, reinforcing the potential of neural networks in predicting water quality and supporting data-driven environmental decision-making