Hybrid Modeling for Nitrate Concentration Prediction Using Linear Regression, Markov Chain, and Artificial Neural Networks

Source abstract

Nitrate pollution is a critical environmental and public health issue worldwide. Our objective is to assess water quality in the Oued Inaouene watershed. To this end, a hybrid approach combining three models was introduced: Linear Regression, Markov chain, and Artificial Neural Network to improve the accuracy of nitrate concentration forecasts. The performance of this model was evaluated using performance measures such as MAE, MSE, d, and R2. The results show a significant improvement when artificial neural networks were introduced into the hybrid approach.