Catalogo Articoli (Spogli Riviste)

OPAC HELP

Titolo:
Interactive initialization of the multilayer perceptron
Autore:
Raudys, A;
Indirizzi:
Inst Math & Informat, LT-2600 Vilnius, Lithuania Inst Math & Informat Vilnius Lithuania LT-2600 T-2600 Vilnius, Lithuania
Titolo Testata:
PATTERN RECOGNITION LETTERS
fascicolo: 10, volume: 21, anno: 2000,
pagine: 907 - 916
SICI:
0167-8655(200009)21:10<907:IIOTMP>2.0.ZU;2-7
Fonte:
ISI
Lingua:
ENG
Soggetto:
NETWORKS;
Keywords:
multilayer perceptron (MLP); training; initialization; data transformation; feature mapping; principal components; Foley-Sammon mapping;
Tipo documento:
Article
Natura:
Periodico
Settore Disciplinare:
Engineering, Computing & Technology
Citazioni:
26
Recensione:
Indirizzi per estratti:
Indirizzo: Raudys, A Inst Math & Informat, Akademijos 4, LT-2600 Vilnius, Lithuania Inst Math & Informat Akademijos 4 Vilnius Lithuania LT-2600 nia
Citazione:
A. Raudys, "Interactive initialization of the multilayer perceptron", PATT REC L, 21(10), 2000, pp. 907-916

Abstract

A new multilayer preceptor initialization method is proposed and compared experimentally with a traditional random initialization method. An operatormaps training-set vectors into a two-variate space, inspects hi-variate training-set vectors and controls the complexity of the decision boundary. Simulations with sixteen real-world pattern classification tasks have shown that in small-scale pattern classification problems, often complex classification rules and non-linear decision boundaries are not necessary. However, in cases where non-linear decision boundaries are required, the proposed weight initialization method is useful. (C) 2000 Elsevier Science B.V. All rights reserved.

ASDD Area Sistemi Dipartimentali e Documentali, Università di Bologna, Catalogo delle riviste ed altri periodici
Documento generato il 04/07/20 alle ore 14:11:34