Wastewater effluent prediction based on fuzzy-rough sets RBF neural networks

Wastewater effluent prediction is very important in wastewater treatment. The process of wastewater treatment is complicated and nonlinear. This paper combines fuzzy-rough sets method with RBF neural networks to predict wastewater important outputs including BOD and COD. In order to select important influence data and reduce influence noise, fuzzy-rough sets method is used to select the influence core data, then to reduce the dimension of data and pre-process samples. A new fuzzy RBFNN was designed to do the prediction test. Through simulation of real wastewater plant data, the method presented is proved to be effective and meaningful.