A constructive algorithm for wavelet neural networks

Authors:
Jinhua Xu;Daniel W. C. Ho
Affiliations:
Institute of System Science, East China Normal University;Department of Mathematics, City University of Hong Kong
Venue:
ICNC'05 Proceedings of the First international conference on Advances in Natural Computation - Volume Part I
Year:
2005

Citing 10
Cited 3

The cascade-correlation learning architecture

Advances in neural information processing systems 2
A resource-allocating network for function interpolation

Neural Computation
Ten lectures on wavelets

Ten lectures on wavelets
System impulse response identification using a multiresolution neural network

Automatica (Journal of IFAC)
On-line Successive Synthesis of Wavelet Networks

Neural Processing Letters
Wavelet neural networks for function learning

IEEE Transactions on Signal Processing
Using wavelet network in nonparametric estimation

IEEE Transactions on Neural Networks
Constructive algorithms for structure learning in feedforward neural networks for regression problems

IEEE Transactions on Neural Networks
Objective functions for training new hidden units in constructive neural networks

IEEE Transactions on Neural Networks
Use of a quasi-Newton method in a feedforward neural network construction algorithm

IEEE Transactions on Neural Networks

Learning algorithm of wavelet network based on sampling theory

Neurocomputing
Identification of the Listeria monocytogenes survival curves in UHT whole milk utilising local linear wavelet neural networks

Expert Systems with Applications: An International Journal
Wavelet neural networks: A practical guide

Neural Networks

Quantified Score

Hi-index	0.00

Visualization

Abstract

In this paper, a new constructive algorithm for wavelet neural networks (WNN) is proposed. Employing the time-frequency localization property of wavelet, the wavelet network is constructed from the low resolution to the high resolution. At each resolution, a new wavelet is initialized as a member of wavelet frames. The input weight freezing technique is used and the Levenberg-Marquardt (LM) algorithm, a quasi-Newton method, is used to train the new wavelet in the WNN. After training, the new wavelet will be added to the wavelet network if the reduction of the residual error between the desired output and WNN output is greater than a threshold. The proposed algorithm is suitable to situations when the wavelet library is very large. The simulations demonstrate the effectiveness of the proposed approach.