| 一种用于非线性函数逼近的小波神经网络 |
| |
| 引用本文： | 许慧,申东日,陈义俊.一种用于非线性函数逼近的小波神经网络[J].自动化与仪器仪表,2003(6):4-6. |
| |
| 作者姓名： | 许慧 申东日 陈义俊 |
| |
| 作者单位： | 辽宁石油化工大学信息工程学院,抚顺,113001 |
| |
| 摘 要： | 提出一种用于非线性函数逼近的小波神经网络，给出了网络的参数训练方法。从信息熵的概念出发，改进了网络参数训练的目标函数，并利用引入动量项的最速下降法训练网络权值、尺度因子和平移因子。仿真实验表明，该小波神经网络用于非线性函数逼近时优于同等规模的BP网络，且其训练方法亦具有收敛速度快、逼近精度高等优点。
|
| 关 键 词： | 小波分析 小波神经网络 信息熵 函数逼近 BP网络 非线性系统 |
| 文章编号： | 1001-9227(2003)06-0004-03 |
A kind of wavelet neural networks used in approaching non- linear functions |
| |
| Xu Hui,etc..A kind of wavelet neural networks used in approaching non- linear functions[J].Automation & Instrumentation,2003(6):4-6. |
| |
| Authors: | Xu Hui etc |
| |
| Abstract: | A kind of wavelet neural networks used in approaching non-linear functions is proposed,and the method of parameter study of the network is given.By the conception of information entropy,we proposed a new improved objective function,the weights of the network,scale factor and displacement factor are studied by the steepest descent method.The results of simulation indicate that the wavelet neural networks algorithm is better than the BP neural networks when used in approaching non-linear functions.The studied method is fast in its convergence speed and has a good approaching precision. |
| |
| Keywords: | Wavelet analysis Wavelet neural networks Information entropy Function approach |
| 本文献已被 CNKI 维普 万方数据 等数据库收录！ |
