Application of artificial neural network (ANN) for water quality index (WQI) prediction for the river Warta, Poland

Abstract The aim of this paper is to present the potential using neural network modelling for prediction surface water quality index (WQI). An artificial has been performed physicochemical parameters (TDS, chloride, TH, nitrate, and manganese) as an input layer model, WQI output layer. have taken from five measuring stations river Warta in years 2014-2018 via Chief Inspectorate Environmental Protection (GIOŚ). best results were obtained networks with 5 neurons hidden A high correlation coefficient (general within subsets) 0.9792, low level MSE each subset (training, test, validation), well RMSE at a 0.624507639 serve confirmation. Additionally, maximum percentage error value did not exceed 4%, which confirms conformity real data comparison those during prediction. aforementioned clearly that ANN models are effective Surface may be regarded adequate application simulation by units monitoring condition environment.