任意非线性函数的神经网络实现及其仿真分析

The Simulation Analysis of Realization of Any Nonlinear Function By Using Neural Network

本文采用计算机仿真的方法,对多层神经网络描述任意有界非线性函数的特性进行了分析和实验验证,并对网络的隐层数和每一隐层的神经元个数给网络输出带来的影响进行了分析,同时给出了仿真结果。

In this paper, the property that a multi-layered neural network can describe any bounded nonlinear function is analyzed and verified by simulation experiment. The effect of the number of hidden layer neurons on network output is also analyzed, and the simulation results are given.