Fuzzy Neural Network Model and Its Application in Water Quality Evaluation

In view of the defect of traditional water quality evaluation model, based on fuzzy neural network theory, a new model of fuzzy neural network (FNN) comprehensive evaluation is developed to evaluate surface water quality in Suzhou. Fuzzy neural network is a new type neural network consisting radical basis network and compete neural network, which is simple in structure, easy for training and wide used. FNN model is applied to evaluate water quality at representative sections in Suzhou surface area from the year 1999-2002. The results indicate that FNN model is suitable for water quality evaluation. By analysis, it is important to pay attention to bring into effective measures for pollution control.

I. Introduction

Water quality evaluation is to distinguish the grade of water quality according to water quality standard. This is a typical problem on pattern recognition. In the domain of pattern recognition, artificial neural network has shown good characters. Traditional water quality evaluation method usually adopt accurate mathematic model to describe. But there are many factors influencing water quality, the relationship between evaluation indicators and standards is nonlinear. The relationship among every grade in evaluation standard is fuzzy and grey. There are much difficulties if adopting certain and uncertain method. Artificial neural network (ANN) has the ability of learning and repressing random nonlinear relationship. Artificial neural network can overcome actual difficulties of traditional water quality evaluation method when ANN is used to evaluate water quality.