摘要
运用BP(误差反向传播)模型对水环境质量进行综合评价,主要针对以前BP模型在水质评价中存在的学习训练样本过少,没有检验样本等问题,用随机数发生器在每个级别范围内产生大量的数据作为训练样本和检验样本,并尝试以MSE函数生成均方误差作为检验样本的输出值与期望输出值的比较,检验网络评价未知样本的能力,大大提高了神经网络评价水质时的精度。
This paper presents an error back propagation model to evaluate water quality comprehensively.Aiming at the problems of few training samples and no testing samples when error back propagation model is applied in the field of water quality assessment,a great number of data can be generated through uniform distribution random function as training samples and testing samples.Trying to anlysize the relationships of the output of testing samples and the expecting output using the MSE function in MATLAB 7 could raise the accountability of new sampling.This can greatly improve the accurancy of the ANN model in the water quality evaluation.
出处
《环境工程》
CAS
CSCD
北大核心
2007年第1期69-71,共3页
Environmental Engineering
基金
国家自然科学基金资助项目(50378008)
关键词
BP神经网络
水质综合评价
随机数样本
检验
BPANN,water quality synthetic evaluation, random samples and testing
