UNIVERSAL APPROXIMATION IN P-MEAN BY NEURAL NETWORKS

Authors
BURTON RM DEHLING HG
Citation
Rm. Burton et Hg. Dehling, UNIVERSAL APPROXIMATION IN P-MEAN BY NEURAL NETWORKS, Neural networks, 11(4), 1998, pp. 661-667
Citations number
7
Categorie Soggetti
Computer Science Artificial Intelligence","Computer Science Artificial Intelligence
Journal title
Neural networks → ACNP
ISSN journal
08936080
Volume
11
Issue
4
Year of publication
1998
Pages
661 - 667
Database
ISI
SICI code
0893-6080(1998)11:4<661:UAIPBN>2.0.ZU;2-O
Abstract
A feedforward neural net with d input neurons and with a single hidden layer of n neurons is given by [GRAPHICS] where a(j), theta(j), w(ji) is an element of R. In this paper we study the approximation of arbit rary functions f: R-d --> R by a neural net in an L-p(mu) norm for som e finite measure mu on R-d. We prove that under natural moment conditi ons, a neural net with non-polynomial function can approximate any giv en function. (C) 1998 Elsevier Science Ltd. All rights reserved.