BP原理及其在林木胸径模拟中的实现  

Principle of BP Neural Network and Its Simulation in Diameter at Breast Height

作　　者：车少辉[1] 张建国[1] 段爱国[1] 骆昱春[2] 

机构地区：[1]中国林业科学研究院林业研究所,北京100091 [2]江西省林业科学院

出　　处：《东北林业大学学报》2011年第8期116-119,共4页Journal of Northeast Forestry University

摘　　要：基于对神经元模型、网络拓扑结构和学习算法3要素的描述,系统地剖析了一个BP3-2-1结构的模型模拟非线性函数的一般过程。基于杉木人工林密度实验数据,以初始密度、立地指数和林龄为输入,林分平均胸径为输出,利用MATLAB(R2010b)神经网络工具箱创建、训练BP模型,模型的均方误差Ems=0.01,实测值和预测值之间的相关系数R=0.977。The aim of this paper is to systematically reveal the principle of BP model with 3-2-1 structure according to the introduction of neuron model, topology and back-propagation algorithm for supervised learning. A BP model is set up in the MATLAB （R2010b） environment to simulate the mean diameter at breast height of a Cunninghamia lanceolata plantation, using initial density, site index and stand age as input parameters. The mean square error of the model is 0.01 and the correlation coefficient between the independent variable and the dependent variable is 0. 977.

关 键 词：BP神经元模型 网络拓扑结构 学习算法 胸径模拟 

分 类 号：S757[农业科学—森林经理学]