The hidden neurons selection of the wavelet networks using support vector machines and ridge regression

被引：25

作者：

Han, Min ^{[1
]}

Yin, Jia ^{[1
]}

机构：

[1] Dalian Univ Technol, Sch Elect & Informat Engn, Dalian 116023, Peoples R China

来源：

NEUROCOMPUTING | 2008年 / 72卷 / 1-3期

关键词：

Wavelet network; Support vector machine; Hidden neurons selection; Ridge regression;

D O I：

10.1016/j.neucom.2007.12.009

中图分类号：

TP18 [人工智能理论];

学科分类号：

081104 ; 0812 ; 0835 ; 1405 ;

摘要：

A 1-norm support vector machine stepwise (SVMS) algorithm is proposed for the hidden neurons selection of wavelet networks (WNs). In this new algorithm, the linear programming support vector machine (LPSVM) is employed to pre-select the hidden neurons, and then a stepwise selection algorithm based on ridge regression is introduced to select hidden neurons from the pre-selection. The main advantages of the new algorithm are that it can get rid of the influence of the ill conditioning of the matrix and deal with the problems that involve a great number of candidate neurons or a large size of samples. Four examples are provided to illustrate the efficiency of the new algorithm. (c) 2007 Elsevier B.V. All rights reserved.

引用

页码：471 / 479

页数：9

共 22 条

[1] System identification via optimised wavelet-based neural networks [J].

Alonge, F ;

D'Ippolito, F ;

Raimondi, FM .

IEE PROCEEDINGS-CONTROL THEORY AND APPLICATIONS, 2003, 150 (02) :147-154

[2] USING MUTUAL INFORMATION FOR SELECTING FEATURES IN SUPERVISED NEURAL-NET LEARNING [J].

BATTITI, R .

IEEE TRANSACTIONS ON NEURAL NETWORKS, 1994, 5 (04) :537-550

[3] A new class of wavelet networks for nonlinear system identification [J].

Billings, SA ;

Wei, HL .

IEEE TRANSACTIONS ON NEURAL NETWORKS, 2005, 16 (04) :862-874

[4] The wavelet-NARMAX representation: A hybrid model structure combining polynomial models with multiresolution wavelet decompositions [J].

Billings, SA ;

Wei, HL .

INTERNATIONAL JOURNAL OF SYSTEMS SCIENCE, 2005, 36 (03) :137-152

[5]

Bradley P. S., 1998, Machine Learning. Proceedings of the Fifteenth International Conference (ICML'98), P82

[6] Predicting chaotic time series with wavelet networks [J].

1600, Elsevier Science B.V., Amsterdam, Netherlands (85) :1-2

[7] ORTHOGONAL LEAST-SQUARES METHODS AND THEIR APPLICATION TO NON-LINEAR SYSTEM-IDENTIFICATION [J].

CHEN, S ;

BILLINGS, SA ;

LUO, W .

INTERNATIONAL JOURNAL OF CONTROL, 1989, 50 (05) :1873-1896

[8] A modified Gram-Schmidt algorithm with iterative orthogonalization and column pivoting [J].

Dax, A .

LINEAR ALGEBRA AND ITS APPLICATIONS, 2000, 310 (1-3) :25-42

[9] A feature selection Newton method for support vector machine classification [J].

Fung, GM ;

Mangasarian, OL .

COMPUTATIONAL OPTIMIZATION AND APPLICATIONS, 2004, 28 (02) :185-202

[10] Order and delay selection for neural network modelling by identification of linearized models [J].

Gomm, JB ;

Yu, DL .

INTERNATIONAL JOURNAL OF SYSTEMS SCIENCE, 2000, 31 (10) :1273-1283

← 1 2 3 →