重金属复合污染对土壤活性酶影响的神经网络模型初探  

The primary discussion of neural network model for the effect of heavy metal's complex pollution on the activity of soil enzyme

作　　者：李永青[1] 方振东[1] 郭伟[1] 

机构地区：[1]解放军后勤工程学院,重庆400016

出　　处：《重庆工业高等专科学校学报》2003年第1期36-39,共4页Journal of Chongqing Polytechnic College

摘　　要：在环境领域中 ,重金属对土壤的污染可通过常规的多元回归方法进行研究。但由于这一方法需附加假设条件、易出现共线性等缺点 ,而使结果偏离实际。本文建立神经网络的两个模型 ,寻找了潮褐土中Cd、Zn、Pb的复合污染与对 4种土壤酶活性影响之间的关系 ,并分析了各重金属对土壤酶活性的抑制效应 ,为实际预测提供了依据。结果表明神经网络的方法简单易行 。In field of environment, soil pollution resulting from heavy metal usually is studied through linear regress method. However, because this method needs assumptive condition and easily forms conformal linearity and so on, the results may deviate from fact. The article set up two neural network models to find out relations between complex pollution of Cd, Zn, Pb and the effect on the four activity of soil enzyme in damp-brownness soil. At the same time they analyze the prohibitive effect of each heavy metal on the activity of soil enzyme, offer foundation for forecast in fact. The results show that neural network method is simple and more easily operated than linear regress method.

关 键 词：神经网络 重金属 复合污染 土壤酶活性 

分 类 号：X833[环境科学与工程—环境工程]