Author :
Zhang, Qinghua ; Benveniste, Albert
Author_Institution :
Linkopings Tekniska Hogskola, Sweden
fDate :
11/1/1992 12:00:00 AM
Abstract :
A wavelet network concept, which is based on wavelet transform theory, is proposed as an alternative to feedforward neural networks for approximating arbitrary nonlinear functions. The basic idea is to replace the neurons by `wavelons´, i.e., computing units obtained by cascading an affine transform and a multidimensional wavelet. Then these affine transforms and the synaptic weights must be identified from possibly noise corrupted input/output data. An algorithm of backpropagation type is proposed for wavelet network training, and experimental results are reported
Keywords :
backpropagation; feedforward neural nets; wavelet transforms; affine transform; arbitrary nonlinear functions approximation; backpropagation; feedforward neural networks; learning algorithm; multidimensional wavelet; synaptic weights; wavelet network; wavelet transform theory; Backpropagation algorithms; Continuous wavelet transforms; Convergence; Discrete transforms; Feedforward neural networks; Helium; Neural networks; Nonlinear systems; Power system modeling; Wavelet transforms;
Journal_Title :
Neural Networks, IEEE Transactions on
