Self-construction algorithm for synthesis of wavelet networks

被引：12

作者：

Kan, K ^{[1
]}

Wong, K ^{[1
]}

机构：

[1] City Univ Hong Kong, Dept Elect Engn, Tat Chee Ave, Hong Kong, Peoples R China

来源：

ELECTRONICS LETTERS | 1998年 / 34卷 / 20期

关键词：

D O I：

10.1049/el:19981364

中图分类号：

TM [电工技术]; TN [电子技术、通信技术];

学科分类号：

0808 ; 0809 ;

摘要：

A novel algorithm for the synthesis of wavelet networks is proposed which possesses the self-construction capability. It is referred as recursive variance suppression growth. Using this method, the network is allowed to start with a null hidden neuron and then grows autonomously. Simulations show that the proposed method outperforms other reported methods.

引用

页码：1953 / 1955

页数：3

共 4 条

[1] COMPETITIVE LEARNING ALGORITHMS FOR VECTOR QUANTIZATION [J].

AHALT, SC ;

KRISHNAMURTHY, AK ;

CHEN, PK ;

MELTON, DE .

NEURAL NETWORKS, 1990, 3 (03) :277-290

[2] NEURAL NETWORK ADAPTIVE WAVELETS FOR SIGNAL REPRESENTATION AND CLASSIFICATION [J].

SZU, HH ;

TELFER, B ;

KADAMBE, S .

OPTICAL ENGINEERING, 1992, 31 (09) :1907-1916

[3] RIVAL PENALIZED COMPETITIVE LEARNING FOR CLUSTERING ANALYSIS, RBF NET, AND CURVE DETECTION [J].

XU, L ;

KRZYZAK, A ;

OJA, E .

IEEE TRANSACTIONS ON NEURAL NETWORKS, 1993, 4 (04) :636-649

[4] WAVELET NETWORKS [J].

ZHANG, QG ;

BENVENISTE, A .

IEEE TRANSACTIONS ON NEURAL NETWORKS, 1992, 3 (06) :889-898

← 1 →