Presentation 1995/11/17
CONSTRUCTIVE FUNCTION-APPROXIMATIONS BY THREE-LAYER ARTIFICIAL NEURAL NETWORKS AND SATURATIONS OF THE APPROXIMATIONS
Shin SUZUKI,
PDF Download Page PDF download Page Link
Abstract(in Japanese) (See Japanese page)
Abstract(in English) This paper presents constructive function-approximations by three-layer artificial neural networks with (1) trigonometric, (2) piecewise linear, and (3) sigmoidal hidden-layer units. The approximations provide (a) approximating-network equations, (b) specifications for the numbers of hidden-layer units, (c) approximation error estimations, and (d) saturations of the approximations. These results can easily be applied to non-periodic functions defined on a bounded subset.
Keyword(in Japanese) (See Japanese page)
Keyword(in English) Artificial neural networks / Function approximation / Network equation / Number of hidden-layer units / Estimation of approximation error / Saturation of approximation
Paper # NLP95-71
Date of Issue

Conference Information
Committee NLP
Conference Date 1995/11/17(1days)
Place (in Japanese) (See Japanese page)
Place (in English)
Topics (in Japanese) (See Japanese page)
Topics (in English)
Chair
Vice Chair
Secretary
Assistant

Paper Information
Registration To Nonlinear Problems (NLP)
Language JPN
Title (in Japanese) (See Japanese page)
Sub Title (in Japanese) (See Japanese page)
Title (in English) CONSTRUCTIVE FUNCTION-APPROXIMATIONS BY THREE-LAYER ARTIFICIAL NEURAL NETWORKS AND SATURATIONS OF THE APPROXIMATIONS
Sub Title (in English)
Keyword(1) Artificial neural networks
Keyword(2) Function approximation
Keyword(3) Network equation
Keyword(4) Number of hidden-layer units
Keyword(5) Estimation of approximation error
Keyword(6) Saturation of approximation
1st Author's Name Shin SUZUKI
1st Author's Affiliation NTT Basic Research Laboratories()
Date 1995/11/17
Paper # NLP95-71
Volume (vol) vol.95
Number (no) 368
Page pp.pp.-
#Pages 8
Date of Issue