A new SOM Initialization Algorithm for Nonvectorial Data

Salvatore Gaglio, Antonino Fiannaca, Alfonso Urso, Salvatore Gaglio, Riccardo Rizzo, Antonino Fiannaca

Risultato della ricerca: Article

2 Citazioni (Scopus)

Abstract

Self Organizing Maps (SOMs) are widely used mapping and clustering algorithms family. It is also well known that the performances of the maps in terms of quality of result and learning speed are strongly dependent from the neuron weights initialization. This drawback is common to all the SOM algorithms, and critical for a new SOM algorithm, the Median SOM (M-SOM), developed in order to map datasets characterized by a dissimilarity matrix. In this paper an initialization technique of M-SOM is proposed and compared to the initialization techniques proposed in the original paper. The results show that the proposed initialization technique assures faster learning and better performance in terms of quantization error.
Lingua originaleEnglish
pagine (da-a)41-48
Numero di pagine8
RivistaLECTURE NOTES IN ARTIFICIAL INTELLIGENCE
Volume5177
Stato di pubblicazionePublished - 2008

Fingerprint

Self organizing maps
Self-organizing Map
Initialization
Dissimilarity
Clustering algorithms
Neurons
Clustering Algorithm
Neuron
Quantization
Dependent
Learning

All Science Journal Classification (ASJC) codes

  • Theoretical Computer Science
  • Computer Science(all)

Cita questo

A new SOM Initialization Algorithm for Nonvectorial Data. / Gaglio, Salvatore; Fiannaca, Antonino; Urso, Alfonso; Gaglio, Salvatore; Rizzo, Riccardo; Fiannaca, Antonino.

In: LECTURE NOTES IN ARTIFICIAL INTELLIGENCE, Vol. 5177, 2008, pag. 41-48.

Risultato della ricerca: Article

Gaglio, S, Fiannaca, A, Urso, A, Gaglio, S, Rizzo, R & Fiannaca, A 2008, 'A new SOM Initialization Algorithm for Nonvectorial Data', LECTURE NOTES IN ARTIFICIAL INTELLIGENCE, vol. 5177, pagg. 41-48.
Gaglio, Salvatore ; Fiannaca, Antonino ; Urso, Alfonso ; Gaglio, Salvatore ; Rizzo, Riccardo ; Fiannaca, Antonino. / A new SOM Initialization Algorithm for Nonvectorial Data. In: LECTURE NOTES IN ARTIFICIAL INTELLIGENCE. 2008 ; Vol. 5177. pagg. 41-48.
@article{d7d9e78a1c7547f38c104b0a625a8659,
title = "A new SOM Initialization Algorithm for Nonvectorial Data",
abstract = "Self Organizing Maps (SOMs) are widely used mapping and clustering algorithms family. It is also well known that the performances of the maps in terms of quality of result and learning speed are strongly dependent from the neuron weights initialization. This drawback is common to all the SOM algorithms, and critical for a new SOM algorithm, the Median SOM (M-SOM), developed in order to map datasets characterized by a dissimilarity matrix. In this paper an initialization technique of M-SOM is proposed and compared to the initialization techniques proposed in the original paper. The results show that the proposed initialization technique assures faster learning and better performance in terms of quantization error.",
keywords = "Median SOM, initialization, pairwise data",
author = "Salvatore Gaglio and Antonino Fiannaca and Alfonso Urso and Salvatore Gaglio and Riccardo Rizzo and Antonino Fiannaca",
year = "2008",
language = "English",
volume = "5177",
pages = "41--48",
journal = "LECTURE NOTES IN ARTIFICIAL INTELLIGENCE",

}

TY - JOUR

T1 - A new SOM Initialization Algorithm for Nonvectorial Data

AU - Gaglio, Salvatore

AU - Fiannaca, Antonino

AU - Urso, Alfonso

AU - Gaglio, Salvatore

AU - Rizzo, Riccardo

AU - Fiannaca, Antonino

PY - 2008

Y1 - 2008

N2 - Self Organizing Maps (SOMs) are widely used mapping and clustering algorithms family. It is also well known that the performances of the maps in terms of quality of result and learning speed are strongly dependent from the neuron weights initialization. This drawback is common to all the SOM algorithms, and critical for a new SOM algorithm, the Median SOM (M-SOM), developed in order to map datasets characterized by a dissimilarity matrix. In this paper an initialization technique of M-SOM is proposed and compared to the initialization techniques proposed in the original paper. The results show that the proposed initialization technique assures faster learning and better performance in terms of quantization error.

AB - Self Organizing Maps (SOMs) are widely used mapping and clustering algorithms family. It is also well known that the performances of the maps in terms of quality of result and learning speed are strongly dependent from the neuron weights initialization. This drawback is common to all the SOM algorithms, and critical for a new SOM algorithm, the Median SOM (M-SOM), developed in order to map datasets characterized by a dissimilarity matrix. In this paper an initialization technique of M-SOM is proposed and compared to the initialization techniques proposed in the original paper. The results show that the proposed initialization technique assures faster learning and better performance in terms of quantization error.

KW - Median SOM, initialization, pairwise data

UR - http://hdl.handle.net/10447/58930

M3 - Article

VL - 5177

SP - 41

EP - 48

JO - LECTURE NOTES IN ARTIFICIAL INTELLIGENCE

JF - LECTURE NOTES IN ARTIFICIAL INTELLIGENCE

ER -