Application of a Novel Method for Predicting Water Quality Using the Pi-Sigma Neural Network: Gelevera Creek (Giresun, Türkiye)

Abstract

Identifying and measuring potential sources of pollution is essential for water management and pollution control. Using a range of artificial intelligence models to analyze water quality (WQ) is one of the most effective techniques for estimating WQI. In this context, water samples have been collected in monthly from 8 stations of Gelevera Creek. The traditional evaluation with WQI of Gelevera Creek was calculated as average so good WQ. The novel application which is the Single multiplicative neuron (SMN) model, multilayer perceptron and pi-sigma artificial neural networks (PS-ANNs) are applied for predicting of WQI. SMN model and PS-ANNs are firstly used for modelling of WQ in the literature. It is noted that the best results of Gelevera Creek were obtained with the PS-ANN. As a result of, it is suggested to obtain the WQI with the proposed optimum PS-ANN instead of using calculation methods such as WQI that include long calculations.