Function approximation using robust wavelet neural networks

Sheng-Tun Li, Shu Ching Chen

Research output: Contribution to journalConference article

28 Citations (Scopus)

Abstract

Wavelet neural networks (WNN) have recently attracted great interest, because of their advantages over radial basis function networks (RBFN) as they are universal approximators but achieve faster convergence and are capable of dealing with the so-called "curse of dimensionality." In addition, WNN are generalized RBFN. However, the generalization performance of WNN trained by least-squares approach deteriorates when outliers are present. In this paper, we propose a robust wavelet neural network based on the theory of robust regression for dealing with outliers in the framework of function approximation. By adaptively adjusting the number of training data involved during training, the efficiency loss in the presence of Gaussian noise is accommodated. Simulation results are demonstrated to validate the generalization ability and efficiency of the proposed network.

Original languageEnglish
Pages (from-to)483-488
Number of pages6
JournalProceedings of the International Conference on Tools with Artificial Intelligence
Publication statusPublished - 2002 Dec 1
Event14th International Conference on Tools with Artificial Intelligence - Washington, DC, United States
Duration: 2002 Jun 42002 Nov 6

Fingerprint

Neural networks
Radial basis function networks

All Science Journal Classification (ASJC) codes

  • Software

Cite this

@article{58ca8465879447208775368c9c69441c,
title = "Function approximation using robust wavelet neural networks",
abstract = "Wavelet neural networks (WNN) have recently attracted great interest, because of their advantages over radial basis function networks (RBFN) as they are universal approximators but achieve faster convergence and are capable of dealing with the so-called {"}curse of dimensionality.{"} In addition, WNN are generalized RBFN. However, the generalization performance of WNN trained by least-squares approach deteriorates when outliers are present. In this paper, we propose a robust wavelet neural network based on the theory of robust regression for dealing with outliers in the framework of function approximation. By adaptively adjusting the number of training data involved during training, the efficiency loss in the presence of Gaussian noise is accommodated. Simulation results are demonstrated to validate the generalization ability and efficiency of the proposed network.",
author = "Sheng-Tun Li and Chen, {Shu Ching}",
year = "2002",
month = "12",
day = "1",
language = "English",
pages = "483--488",
journal = "Proceedings of the International Conference on Tools with Artificial Intelligence",
issn = "1063-6730",
publisher = "Institute of Electrical and Electronics Engineers Inc.",

}

Function approximation using robust wavelet neural networks. / Li, Sheng-Tun; Chen, Shu Ching.

In: Proceedings of the International Conference on Tools with Artificial Intelligence, 01.12.2002, p. 483-488.

Research output: Contribution to journalConference article

TY - JOUR

T1 - Function approximation using robust wavelet neural networks

AU - Li, Sheng-Tun

AU - Chen, Shu Ching

PY - 2002/12/1

Y1 - 2002/12/1

N2 - Wavelet neural networks (WNN) have recently attracted great interest, because of their advantages over radial basis function networks (RBFN) as they are universal approximators but achieve faster convergence and are capable of dealing with the so-called "curse of dimensionality." In addition, WNN are generalized RBFN. However, the generalization performance of WNN trained by least-squares approach deteriorates when outliers are present. In this paper, we propose a robust wavelet neural network based on the theory of robust regression for dealing with outliers in the framework of function approximation. By adaptively adjusting the number of training data involved during training, the efficiency loss in the presence of Gaussian noise is accommodated. Simulation results are demonstrated to validate the generalization ability and efficiency of the proposed network.

AB - Wavelet neural networks (WNN) have recently attracted great interest, because of their advantages over radial basis function networks (RBFN) as they are universal approximators but achieve faster convergence and are capable of dealing with the so-called "curse of dimensionality." In addition, WNN are generalized RBFN. However, the generalization performance of WNN trained by least-squares approach deteriorates when outliers are present. In this paper, we propose a robust wavelet neural network based on the theory of robust regression for dealing with outliers in the framework of function approximation. By adaptively adjusting the number of training data involved during training, the efficiency loss in the presence of Gaussian noise is accommodated. Simulation results are demonstrated to validate the generalization ability and efficiency of the proposed network.

UR - http://www.scopus.com/inward/record.url?scp=0036929141&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=0036929141&partnerID=8YFLogxK

M3 - Conference article

AN - SCOPUS:0036929141

SP - 483

EP - 488

JO - Proceedings of the International Conference on Tools with Artificial Intelligence

JF - Proceedings of the International Conference on Tools with Artificial Intelligence

SN - 1063-6730

ER -