A PROCEDURE FOR DETERMINING THE TOPOLOGY OF MULTILAYER FEEDFORWARD NEURAL NETWORKS

被引：74

作者：

WANG, ZN ^{[1
]}

DIMASSIMO, C ^{[1
]}

THAM, MT ^{[1
]}

MORRIS, AJ ^{[1
]}

机构：

[1] UNIV NEWCASTLE UPON TYNE,DEPT CHEM & PROC ENGN,NEWCASTLE TYNE NE1 7RU,TYNE & WEAR,ENGLAND

来源：

NEURAL NETWORKS | 1994年 / 7卷 / 02期

关键词：

NEURAL NETWORK TOPOLOGY; CANONICAL FORM; NUMBER OF NEURONS;

D O I：

10.1016/0893-6080(94)90023-X

中图分类号：

TP18 [人工智能理论];

学科分类号：

081104 ; 0812 ; 0835 ; 1405 ;

摘要：

This contribution describes a procedure for determining the optimal topology of a static three-layer neural network. The method is based on a canonical decomposition technique that establishes a link between the number of neurons in each hidden layer and the dimensions of the subspaces of the canonical decomposition. It ensures that the structure is sufficiently complex to characterise the process being modelled by the network, and also the unicity of the approximation. A practical procedure to implement this method is then outlined and applied to both simulated and real process data.

引用

页码：291 / 300

页数：10

共 7 条

[1]

BETTON F, 1991, DYNAMIC ARTIFICIAL N

[2]

DIMASSIMO C, 1989, NONLINEAR ESTIMATION, P1994

[3]

Lippman R. P., 1987, IEEE ASSP MAGAZI APR, P4

[4]

McClelland J. L., 1986, PARALLEL DISTRIBUTED, V1

[5]

Narendra K S, 1990, IEEE Trans Neural Netw, V1, P4, DOI 10.1109/72.80202

[6] MULTILAYER FEEDFORWARD NEURAL NETWORKS - A CANONICAL FORM APPROXIMATION OF NONLINEARITY [J].

WANG, Z ;

THAM, MT ;

MORRIS, AJ .

INTERNATIONAL JOURNAL OF CONTROL, 1992, 56 (03) :655-672

[7]

WILLIS MJ, 1991, P IEEE D, V138

← 1 →