THE STATISTICAL BASIS OF NEURAL-NETWORK ALGORITHMS - THEORY AND APPLICATIONS

Authors
SWAIN J
Citation
J. Swain, THE STATISTICAL BASIS OF NEURAL-NETWORK ALGORITHMS - THEORY AND APPLICATIONS, Nuclear instruments & methods in physics research. Section A, Accelerators, spectrometers, detectors and associated equipment, 389(1-2), 1997, pp. 271-273
Citations number
12
Categorie Soggetti
Nuclear Sciences & Tecnology","Physics, Particles & Fields","Instument & Instrumentation",Spectroscopy
Journal title
Nuclear instruments & methods in physics research. Section A, Accelerators, spectrometers, detectors and associated equipment → ACNP
ISSN journal
01689002
Volume
389
Issue
1-2
Year of publication
1997
Pages
271 - 273
Database
ISI
SICI code
0168-9002(1997)389:1-2<271:TSBONA>2.0.ZU;2-3
Abstract
In a very general framework, neural network and related algorithms are discussed with emphasis on how the use of correlations between variab les can improve on linear cuts. Mathematical results for a large class of neural algorithms are shown indicating the generic importance of h igh-order correlations between input variables. The dangers of the bli nd application of these newer techniques are stressed and suggestions are made for how to use them wisely.