Groundwater Quality Prediction in Upper and Middle Cheliff Plain, Algeria Using Artificial Intelligence

Groundwater studies are very important for the establishment of a quantitative and qualitative management system that ensures the supply and protection of water in water-scarce semi-arid areas like Algeria. However, monitoring and conducting measurements of groundwater quality and quantity are difficult, costly, and time-consuming; therefore, the use of predictive models such as artificial intelligence (AI) is becoming an attractive alternative. For this purpose, three approaches based on machine learning techniques were used to improve the prediction of water quality in the Upper and Middle Cheliff plain in Algeria. In this study, the most dominant parameters of the water quality index (WQI) that were extracted by principal component analysis (PCA) were used in multilayer perceptron neural network (MLPNN), support vector regression (SVR), and decision tree regression (DTR) models. Various combinations of input data were investigated and models were evaluated in terms of prediction performance, using several statistical criteria. Various potential physicochemical water quality variables were considered for the calculation of the water quality index (WQI) in the study area. This work will be helpful to decision-makers and water authorities for sustainable groundwater resource management and planning.