两种小波网络的函数逼近能力比较与应用

The function approximation ability comparison of two wavelet networks and their applications

基于小波分析中小波变换的两个不同角度,分别介绍了两种不同的小波网络,并对这两种小波网络的构成、模型和函数逼近能力进行了讨论,给出了这两种小波网络对一阶梯函数逼近的仿真结果,得出第二种小波网络具有更好的函数逼近能力,最后介绍了这两种小波网络在信号分类器方面的应用,给出了各自学习模型和仿真结果。

In this paper, from two different aspects of wavelet transformations in wavelet analysis, two wavelet networks are induced, and the constructions,models and function approximation abilities of the two wavelet networks are presented. The simulations of a ladder function approximation with the two wavelet networks are done, and the result is that the function approximation ability of the second wavelet network is better than the first. The two wavelet networks used as signal classification are discussed, and the learning models and simulation results are given in the end.