Ensemble of structure-adaptive self-organizing maps for high performance classification

Research output: Contribution to journalArticle

17 Citations (Scopus)

Abstract

Combining multiple models has been recently exploited for the development of reliable neural networks. This paper introduces a structure-adaptive self-organizing map (SOM) which can adapt the structure as well as the weights, and presents a method to improve the performance by combining the multiple maps. The structure-adaptive SOM places the nodes of prototype vectors into the pattern space properly so as to make the decision boundaries as close to the class boundaries as possible. In order to show the performance of the proposed method, experiments with the unconstrained handwritten digit database of Concordia University in Canada have been conducted.

Original languageEnglish
Pages (from-to)103-114
Number of pages12
JournalInformation sciences
Volume123
Issue number1
DOIs
Publication statusPublished - 2000 Jan 1

Fingerprint

Self organizing maps
Self-organizing Map
Ensemble
High Performance
Multiple Models
Neural networks
Digit
Prototype
Neural Networks
Experiments
Vertex of a graph
Experiment
High performance
Self-organizing map

All Science Journal Classification (ASJC) codes

  • Software
  • Control and Systems Engineering
  • Theoretical Computer Science
  • Computer Science Applications
  • Information Systems and Management
  • Artificial Intelligence

Cite this

@article{b9c6eb192fc3441494a938d2b55ef892,
title = "Ensemble of structure-adaptive self-organizing maps for high performance classification",
abstract = "Combining multiple models has been recently exploited for the development of reliable neural networks. This paper introduces a structure-adaptive self-organizing map (SOM) which can adapt the structure as well as the weights, and presents a method to improve the performance by combining the multiple maps. The structure-adaptive SOM places the nodes of prototype vectors into the pattern space properly so as to make the decision boundaries as close to the class boundaries as possible. In order to show the performance of the proposed method, experiments with the unconstrained handwritten digit database of Concordia University in Canada have been conducted.",
author = "Sung-Bae Cho",
year = "2000",
month = "1",
day = "1",
doi = "10.1016/S0020-0255(99)00112-7",
language = "English",
volume = "123",
pages = "103--114",
journal = "Information Sciences",
issn = "0020-0255",
publisher = "Elsevier Inc.",
number = "1",

}

Ensemble of structure-adaptive self-organizing maps for high performance classification. / Cho, Sung-Bae.

In: Information sciences, Vol. 123, No. 1, 01.01.2000, p. 103-114.

Research output: Contribution to journalArticle

TY - JOUR

T1 - Ensemble of structure-adaptive self-organizing maps for high performance classification

AU - Cho, Sung-Bae

PY - 2000/1/1

Y1 - 2000/1/1

N2 - Combining multiple models has been recently exploited for the development of reliable neural networks. This paper introduces a structure-adaptive self-organizing map (SOM) which can adapt the structure as well as the weights, and presents a method to improve the performance by combining the multiple maps. The structure-adaptive SOM places the nodes of prototype vectors into the pattern space properly so as to make the decision boundaries as close to the class boundaries as possible. In order to show the performance of the proposed method, experiments with the unconstrained handwritten digit database of Concordia University in Canada have been conducted.

AB - Combining multiple models has been recently exploited for the development of reliable neural networks. This paper introduces a structure-adaptive self-organizing map (SOM) which can adapt the structure as well as the weights, and presents a method to improve the performance by combining the multiple maps. The structure-adaptive SOM places the nodes of prototype vectors into the pattern space properly so as to make the decision boundaries as close to the class boundaries as possible. In order to show the performance of the proposed method, experiments with the unconstrained handwritten digit database of Concordia University in Canada have been conducted.

UR - http://www.scopus.com/inward/record.url?scp=0033900205&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=0033900205&partnerID=8YFLogxK

U2 - 10.1016/S0020-0255(99)00112-7

DO - 10.1016/S0020-0255(99)00112-7

M3 - Article

AN - SCOPUS:0033900205

VL - 123

SP - 103

EP - 114

JO - Information Sciences

JF - Information Sciences

SN - 0020-0255

IS - 1

ER -