Evaluation of the accuracy multilayer perceptron neural network with sensitive analyze in estimating monthly quality of electrical conductivity parameter in Maroon Basin

. In the present paper, in order to predict the quality electrical conductivity parameters (Ec ) in Maroon basin located in Khuzestan province, a multilayer artificial neural network of perceptron was used. The data of ۱۹۹۰-۲۰۲۱ were used for prediction. For the input and output layer, the linear function was used and for the hidden layer, different stimulus functions were used with the propagation training algorithm. Finally, the model of artificial neural network with structure (۶-۱۲-۱) with hyperbolic tangent stimulus function with coefficients of determination of ۰.۹۲, and root mean squares of ۱.۹۶, respectively, the best optimal models for EC quality parameters were selected