Self-organizing feature map with a momentum term

Research output: Contribution to journalArticle

2 Citations (Scopus)

Abstract

The objectives of this paper are to derive a momentum term in the Kohonen's self-organizing feature map algorithm theoretically and to show the effectiveness of the term by computer simulations. We will derive the self-organizing feature map algorithm having the momentum term through the following assumptions: (1) The cost function is E(n) = Σ(μ)(n)α(n-μ), where E(μ) is the modified Lyapunov function originally proposed by Ritter and Schulten at the μth learning time and Q is the momentum coefficient. (2) The latest weights are assumed in calculating the cost function E(n). According to our simulations, it has shown that the momentum term in the self-organizing feature map can considerably contribute to the acceleration of the convergence.

Original languageEnglish
Pages (from-to)71-81
Number of pages11
JournalNeurocomputing
Volume10
Issue number1
DOIs
Publication statusPublished - 1996 Jan

Fingerprint

Self organizing maps
Momentum
Costs and Cost Analysis
Cost functions
Computer Simulation
Learning
Weights and Measures
Lyapunov functions
Computer simulation

Keywords

  • Momentum term
  • Self-organizing feature map

ASJC Scopus subject areas

  • Artificial Intelligence
  • Cellular and Molecular Neuroscience

Cite this

Self-organizing feature map with a momentum term. / Hagiwara, Masafumi.

In: Neurocomputing, Vol. 10, No. 1, 01.1996, p. 71-81.

Research output: Contribution to journalArticle

@article{22e35a1274744b73adbcf8ef8120eb08,
title = "Self-organizing feature map with a momentum term",
abstract = "The objectives of this paper are to derive a momentum term in the Kohonen's self-organizing feature map algorithm theoretically and to show the effectiveness of the term by computer simulations. We will derive the self-organizing feature map algorithm having the momentum term through the following assumptions: (1) The cost function is E(n) = Σ(μ)(n)α(n-μ), where E(μ) is the modified Lyapunov function originally proposed by Ritter and Schulten at the μth learning time and Q is the momentum coefficient. (2) The latest weights are assumed in calculating the cost function E(n). According to our simulations, it has shown that the momentum term in the self-organizing feature map can considerably contribute to the acceleration of the convergence.",
keywords = "Momentum term, Self-organizing feature map",
author = "Masafumi Hagiwara",
year = "1996",
month = "1",
doi = "10.1016/0925-2312(94)00056-5",
language = "English",
volume = "10",
pages = "71--81",
journal = "Neurocomputing",
issn = "0925-2312",
publisher = "Elsevier",
number = "1",

}

TY - JOUR

T1 - Self-organizing feature map with a momentum term

AU - Hagiwara, Masafumi

PY - 1996/1

Y1 - 1996/1

N2 - The objectives of this paper are to derive a momentum term in the Kohonen's self-organizing feature map algorithm theoretically and to show the effectiveness of the term by computer simulations. We will derive the self-organizing feature map algorithm having the momentum term through the following assumptions: (1) The cost function is E(n) = Σ(μ)(n)α(n-μ), where E(μ) is the modified Lyapunov function originally proposed by Ritter and Schulten at the μth learning time and Q is the momentum coefficient. (2) The latest weights are assumed in calculating the cost function E(n). According to our simulations, it has shown that the momentum term in the self-organizing feature map can considerably contribute to the acceleration of the convergence.

AB - The objectives of this paper are to derive a momentum term in the Kohonen's self-organizing feature map algorithm theoretically and to show the effectiveness of the term by computer simulations. We will derive the self-organizing feature map algorithm having the momentum term through the following assumptions: (1) The cost function is E(n) = Σ(μ)(n)α(n-μ), where E(μ) is the modified Lyapunov function originally proposed by Ritter and Schulten at the μth learning time and Q is the momentum coefficient. (2) The latest weights are assumed in calculating the cost function E(n). According to our simulations, it has shown that the momentum term in the self-organizing feature map can considerably contribute to the acceleration of the convergence.

KW - Momentum term

KW - Self-organizing feature map

UR - http://www.scopus.com/inward/record.url?scp=0030050210&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=0030050210&partnerID=8YFLogxK

U2 - 10.1016/0925-2312(94)00056-5

DO - 10.1016/0925-2312(94)00056-5

M3 - Article

VL - 10

SP - 71

EP - 81

JO - Neurocomputing

JF - Neurocomputing

SN - 0925-2312

IS - 1

ER -