Expanding self-organizing map for data visualization and cluster analysis

Huidong JIN, Wing Ho SHUM, Kwong Sak LEUNG, Man Leung WONG

Research output: Journal PublicationsJournal Article (refereed)Researchpeer-review

44 Citations (Scopus)

Abstract

The Self-Organizing Map (SOM) is a powerful tool in the exploratory phase of data mining. It is capable of projecting high-dimensional data onto a regular, usually 2dimensional grid of neurons with good neighborhood preservation between two spaces. However, due to the dimensional conflict, the neighborhood preservation cannot always lead to perfect topology preservation. In this paper, we establish an Expanding SOM (ESOM) to preserve better topology between the two spaces. Besides the neighborhood relationship, our ESOM can detect and preserve an ordering relationship using an expanding mechanism. The computational complexity of the ESOM is comparable with that of the SOM. Our experiment results demonstrate that the ESOM constructs better mappings than the classic SOM, especially, in terms of the topological error. Furthermore, clustering results generated by the ESOM are more accurate than those obtained by the SOM.
Original languageEnglish
Pages (from-to)157-173
Number of pages17
JournalInformation Sciences
Volume163
Issue number1-3
DOIs
Publication statusPublished - 14 Jun 2004

Fingerprint

Expanding Maps
Data visualization
Data Visualization
Self organizing maps
Cluster analysis
Self-organizing Map
Cluster Analysis
Preservation
Topology Preservation
Topology
High-dimensional Data
Neurons
Data mining
Neuron
Computational complexity
Data Mining
Computational Complexity
Self-organizing map
Clustering
Grid

Cite this

JIN, Huidong ; SHUM, Wing Ho ; LEUNG, Kwong Sak ; WONG, Man Leung. / Expanding self-organizing map for data visualization and cluster analysis. In: Information Sciences. 2004 ; Vol. 163, No. 1-3. pp. 157-173.
@article{f8107b12a29c421ab0f5d51f1aea170b,
title = "Expanding self-organizing map for data visualization and cluster analysis",
abstract = "The Self-Organizing Map (SOM) is a powerful tool in the exploratory phase of data mining. It is capable of projecting high-dimensional data onto a regular, usually 2dimensional grid of neurons with good neighborhood preservation between two spaces. However, due to the dimensional conflict, the neighborhood preservation cannot always lead to perfect topology preservation. In this paper, we establish an Expanding SOM (ESOM) to preserve better topology between the two spaces. Besides the neighborhood relationship, our ESOM can detect and preserve an ordering relationship using an expanding mechanism. The computational complexity of the ESOM is comparable with that of the SOM. Our experiment results demonstrate that the ESOM constructs better mappings than the classic SOM, especially, in terms of the topological error. Furthermore, clustering results generated by the ESOM are more accurate than those obtained by the SOM.",
author = "Huidong JIN and SHUM, {Wing Ho} and LEUNG, {Kwong Sak} and WONG, {Man Leung}",
year = "2004",
month = "6",
day = "14",
doi = "10.1016/j.ins.2003.03.020",
language = "English",
volume = "163",
pages = "157--173",
journal = "Information Sciences",
issn = "0020-0255",
publisher = "Elsevier Inc.",
number = "1-3",

}

Expanding self-organizing map for data visualization and cluster analysis. / JIN, Huidong; SHUM, Wing Ho; LEUNG, Kwong Sak; WONG, Man Leung.

In: Information Sciences, Vol. 163, No. 1-3, 14.06.2004, p. 157-173.

Research output: Journal PublicationsJournal Article (refereed)Researchpeer-review

TY - JOUR

T1 - Expanding self-organizing map for data visualization and cluster analysis

AU - JIN, Huidong

AU - SHUM, Wing Ho

AU - LEUNG, Kwong Sak

AU - WONG, Man Leung

PY - 2004/6/14

Y1 - 2004/6/14

N2 - The Self-Organizing Map (SOM) is a powerful tool in the exploratory phase of data mining. It is capable of projecting high-dimensional data onto a regular, usually 2dimensional grid of neurons with good neighborhood preservation between two spaces. However, due to the dimensional conflict, the neighborhood preservation cannot always lead to perfect topology preservation. In this paper, we establish an Expanding SOM (ESOM) to preserve better topology between the two spaces. Besides the neighborhood relationship, our ESOM can detect and preserve an ordering relationship using an expanding mechanism. The computational complexity of the ESOM is comparable with that of the SOM. Our experiment results demonstrate that the ESOM constructs better mappings than the classic SOM, especially, in terms of the topological error. Furthermore, clustering results generated by the ESOM are more accurate than those obtained by the SOM.

AB - The Self-Organizing Map (SOM) is a powerful tool in the exploratory phase of data mining. It is capable of projecting high-dimensional data onto a regular, usually 2dimensional grid of neurons with good neighborhood preservation between two spaces. However, due to the dimensional conflict, the neighborhood preservation cannot always lead to perfect topology preservation. In this paper, we establish an Expanding SOM (ESOM) to preserve better topology between the two spaces. Besides the neighborhood relationship, our ESOM can detect and preserve an ordering relationship using an expanding mechanism. The computational complexity of the ESOM is comparable with that of the SOM. Our experiment results demonstrate that the ESOM constructs better mappings than the classic SOM, especially, in terms of the topological error. Furthermore, clustering results generated by the ESOM are more accurate than those obtained by the SOM.

UR - http://commons.ln.edu.hk/sw_master/2107

U2 - 10.1016/j.ins.2003.03.020

DO - 10.1016/j.ins.2003.03.020

M3 - Journal Article (refereed)

VL - 163

SP - 157

EP - 173

JO - Information Sciences

JF - Information Sciences

SN - 0020-0255

IS - 1-3

ER -