Modeling of Tabriz Plain Rainfall Using Artificial Neural Networks

Fast spreading of using artificial neural networks (ANNs), black box and qualified models, in different sciences such as hydrology shows its studying necessity and values. The most important applications of the model in hydrology are water quality modeling and prediction, optimization, classification, estimation of hydrological phenomena and parameters. The aim of this paper was to provide application of artificial neural networks, empirical equation of ascertaining hidden nodes and discussing their strengths and limitations for presenting rainfall artificial neural network forecasting model of Tabriz plain area. In this modeling six different structures of ANNs were used in which the feed forward network with six input nodes and a hidden layer composed the best model. This model was used for showing the effects of learning sample and hidden nodes numbers on minimizing of the model error.