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, n
o 4 (14 ref.), pp. 705-719
Langue / Language
Anglais
Editeur / Publisher
World Scientific, Singapore, SINGAPOUR
(1987)
(Revue)
Mots-clés anglais / English Keywords
;
;
;
;
;
;
;
;
;
;
;
Mots-clés français / French Keywords
;
;
;
;
;
;
;
;
;
;
;
Mots-clés espagnols / Spanish Keywords
;
;
;
;
;
;
;
;
;
;
;
Localisation / Location
INIST-CNRS, Cote INIST : 22088, 35400002451152.0040
Nº notice refdoc (ud4) : 3955700
