概率神经网络及其在河流水质评价中的应用研究

摘要

This paper expounded the basic structure and construction of probability nerve net appraisal model. According to the aim of factor choice and the situation of actual river pollution, appraising factors were selected, and river water quality was evaluated by adopting the model of probability nerve net. With the help of well-trained model of probability nerve net, the water quality of two monitoring profiles on Yuncheng section of the Fen Rive was appraised as poor-V, Vand poor-V,for the high-water, mean water and low-water period , respectively. The appraisal results state clearly that the appraisal model of probability nerve net has quite practicability.%阐述了概率神经网络评价模型的基本结构和构建过程.根据因子选择的目的及河流的实际污染情况,确定评价因子,采用概率神经网络(PNN)模型对河流水质进行评价,并且应用训练好的概率神经网络水质评价模型对汾河运城段两监测断面2010年丰水期、平水期、枯水期综合水质评价为劣Ⅴ类、Ⅴ类、劣Ⅴ类.评价结果表明,概率神经网络评价模型具有很好的实用性.