The prediction of comprehensive pollution indexes of Taihu Lake based on BP network

摘要

In order to provide decision support for management platform of Taihu Lake,this paper studied BP(Back Propagation) neural network model and used it into predicting of the comprehensive pollution index of inter rivers and out rivers of Taihu Lake.This paper established a three-layer BP neural network prediction model,based on Taihu routine environmental monitoring data,predicted the quality of water,studied the relationship between the variation of water quality of Taihu and inflow and outflow rivers of Taihu.The prediction of 2005 year Taihu's water quality showed that,in 2005 the water quality pollution is more seriously than before,generally V water quality.The result is in line with the pollution situation development trend of Taihu Lake.The reason why BP neural network was selected are as follows.Compared with the former methods,BP network method have good adaptability,higher precision,better response indexes of water indexes' internal change rules.What this paper studied can provide scientific basis for limiting water environmental pollution.