基于Matlab的BP神经网络结构与函数逼近能力的关系分析  

Analysis of Relation between the Structure of BP Feed-forward Neural Network and Precision of Function Proximate Based on Matlab

作　　者：罗玉春[1] 都洪基[1] 崔芳芳[1] 

机构地区：[1]南京理工大学动力工程学院,江苏南京210094

出　　处：《现代电子技术》2007年第24期88-90,共3页Modern Electronics Technique

摘　　要：人工神经网络是一种非线性动态数学模型,广泛应用于非线性系统建模、系统辨识、函数逼近等方面。介绍BP网络的结构和学习过程,并介绍利用Matlab人工神经网络工具箱设计BP网络的步骤,在此基础上设计了BP网络以验证其函数逼近能力,仿真结果说明了BP网络具有很强的函数逼近能力。并分析BP网络结构和函数逼近能力的关系,得出网络的结构直接影响网络对函数的逼近能力和效果。Artificial Neural Network is a nonlinear dynamic mathematic model and can be applied to model nonlinear system,class identification and function approximation. In this paper,the structure and learning process of BP feed- forward neural network are presented. The paper introduces the approach to design a BP feed - forward neural network using Matlab/neural network toolbox, based on the network verify the ability of function approximation. The result shows that BP feed -forward neural network has a strong ability to approximate function. The paper also analyzes the relation between the structure of BP feed-forward neural network and precision of function approximate, educing that the structure BP feed forward neural network directly aLfeets the approximate ability and effect.

关 键 词：人工神经网络 BP网络 函数逼近 MATLAB 

分 类 号：TP389[自动化与计算机技术—计算机系统结构]