RefDoc
Haut

Faire une nouvelle recherche
Make a new search
Lancer la recherche


Titre du document / Document title

Boosting performance in neural networks

Auteur(s) / Author(s)

DRUCKER H. (1) ; SCHAPIRE R. ; SIMARD P. (1) ;

Affiliation(s) du ou des auteurs / Author(s) Affiliation(s)

(1) AT&T Bell Laboratories, Holmdel NJ 07733, ETATS-UNIS

Résumé / Abstract

A boosting algorithm, based on the probably approximately correct (PAC) learning model is used to construct an ensemble of neural networks that significantly improves performance (compared to a single network) in optical character recognition (OCR) problems. The effect of boosting is reported on four handwritten image databases consisting of 12 000 digits from segmented ZIP Codes from the United States Postal Service and the following from the National Institute of Standards and Technology: 220 000 digits, 45 000 upper case letters, and 45 000 lower case letters. We use to performance measures: the raw error rate (no rejects) and the reject rate required to achieve a 1% error rate on the patterns not rejected. Boosting improved performance significantly, and, in some cases, dramatically

Revue / Journal Title

International journal of pattern recognition and artificial intelligence    ISSN  0218-0014 

Source / Source

1993, vol. 7, no 4 (14 ref.), pp. 705-719

Langue / Language

Anglais

Editeur / Publisher

World Scientific, Singapore, SINGAPOUR  (1987) (Revue)

Mots-clés anglais / English Keywords

Neural network

;

Artificial intelligence

;

Connectionism

;

Algorithm

;

Character recognition

;

Optics

;

Learning

;

Probability

;

Approximation

;

Manuscript document

;

Performance

;

Mots-clés français / French Keywords

Réseau neuronal

;

Intelligence artificielle

;

Connexionnisme

;

Algorithme

;

Reconnaissance caractère

;

Optique

;

Apprentissage

;

Probabilité

;

Approximation

;

Document manuscrit

;

Performance

;

Mots-clés espagnols / Spanish Keywords

Red neuronal

;

Inteligencia artificial

;

Conexionismo

;

Algoritmo

;

Reconocimiento carácter

;

Optica

;

Aprendizaje

;

Probabilidad

;

Aproximacion

;

Documento manuscrito

;

Rendimiento

;

Localisation / Location

INIST-CNRS, Cote INIST : 22088, 35400002451152.0040

Nº notice refdoc (ud4) : 3955700



Faire une nouvelle recherche
Make a new search
Lancer la recherche
Bas