Development of Artificial intelligence model with aid of statistical methods for simulation of water quality indices in north Khartoum area, Sudan

Abstract Groundwater quality evaluation is among the most critical aspects of water management concepts. As a result of overexploitation due to overpopulation, groundwater is severely deteriorating. Water Quality Index (WQI) is commonly used to appraise groundwater quality. In the present study, WQI model is developed to evaluate the groundwater quality in north Khartoum area. To classify water quality for diverse reasons, WQI uses multiple asset classes, sub-indices, and an accumulation function. Multivariate statistical analyses were applied to test the correlation between different variables to draw on the main variables and processes affecting the groundwater quality in the study area. Furthermore, correlation analysis (CA) and principal component analysis (PCA) served as guide for weights assignment in WQI calculations and interpretation. The innovative statistically guided WQI models (SWQI) proved to be superior in groundwater quality assessment. Although SWQI models is powerful and effective tool in groundwater assessment, however, conventional computation is lengthy, time consuming and is often observed with enormous errors during the calculation of sub-indices. In this study, artificial intelligence techniques are applied to cope with these limitations. Multilayer perceptron (MLP) and support vector regression (SVR) models were used for prediction of WQI. The dataset was divided into two groups (ratio 80:20) for training and validation respectively. The predicted models were compared with actual models using four statistical criteria, namely mean square error (MSE), root mean squared Error (RMSE), mean absolute error (MAE), and coefficient of determination (R2). The obtained results revealed the robustness of the artificial intelligence in prediction of WQI in north Khartoum area. The developed approaches in this study shown to be advantageous in groundwater quality assessment.