Strictly modular probabilistic neural networks for pattern recognition

Number of the records: 1

1.

SYSNO ASEP	0411265
Document Type	J - Journal Article
R&D Document Type	Journal Article
Subsidiary J	Ostatní články
Title	Strictly modular probabilistic neural networks for pattern recognition
Author(s)	Grim, Jiří (UTIA-B)_{RID, ORCID} Just, P. (CZ) Pudil, Pavel (UTIA-B)_RID
Source Title	Neural Network World. - : Ústav informatiky AV ČR, v. v. i. - ISSN 1210-0552 Roč. 13, č. 6 (2003), s. 599-615
Number of pages	17 s.
Language	eng - English
Country	CZ - Czech Republic
Keywords	neural networks ; distribution mixtures ; pattern recognition
Subject RIV	BB - Applied Statistics, Operational Research
R&D Projects	GA402/01/0981 GA ČR - Czech Science Foundation (CSF)
CEZ	AV0Z1075907 - UTIA-B
Annotation	Considering the statistical pattern recognition we approximate the unknown class-conditional probability distributions by multivariate Bernoulli mixtures. We show that both the parameter optimization based on EM algorithm and the resulting Bayesian decision-making can be realized by a strictly modular probabilistic neural network. The autonomous adaptation of neurons includes only the locally available information. The properties of the sequential learning procedure are illustrated by numerical examples.
Workplace	Institute of Information Theory and Automation
Contact	Markéta Votavová, votavova@utia.cas.cz, Tel.: 266 052 201.

Number of the records: 1