SOMICA and geometric ICA

Fabian J. Theis; Carlos G. Puntonet; Elmar W. Lang

doi:10.1109/ICIF.2003.177411

SOMICA and geometric ICA

Fabian J. Theis, Carlos G. Puntonet, Elmar W. Lang

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution › peer-review

Abstract

Geometric independent component analysis (ICA) uses a weight update rule that is very similar to the self-organizing map (SOM) learning rule in the case of a trivial neighborhood function. In this paper we use this fact and present a new geometric ICA algorithm that uses a SOM for learning. The separation quality is better in comparison to other geometric algorithms, but the computational cost is higher. Furthermore, this new algorithm wilt provide insight in how to transfer theoretical results from the SOM area to geometric ICA.

Original language	English
Title of host publication	Proceedings of the 6th International Conference on Information Fusion, FUSION 2003
Publisher	IEEE Computer Society
Pages	1457-1464
Number of pages	8
ISBN (Print)	0972184449, 9780972184441
DOIs	https://doi.org/10.1109/ICIF.2003.177411
State	Published - 2003
Externally published	Yes
Event	6th International Conference on Information Fusion, FUSION 2003 - Cairns, QLD, Australia Duration: 8 Jul 2003 → 11 Jul 2003

Publication series

Name	Proceedings of the 6th International Conference on Information Fusion, FUSION 2003
Volume	2

Conference

Conference	6th International Conference on Information Fusion, FUSION 2003
Country/Territory	Australia
City	Cairns, QLD
Period	8/07/03 → 11/07/03

Keywords

Blind source separation
Geometric ICA
Independent component analysis
Self-organizing maps

Access to Document

10.1109/ICIF.2003.177411

Cite this

@inproceedings{cc318b09bf9243388cbfc61283428a8f,

title = "SOMICA and geometric ICA",

abstract = "Geometric independent component analysis (ICA) uses a weight update rule that is very similar to the self-organizing map (SOM) learning rule in the case of a trivial neighborhood function. In this paper we use this fact and present a new geometric ICA algorithm that uses a SOM for learning. The separation quality is better in comparison to other geometric algorithms, but the computational cost is higher. Furthermore, this new algorithm wilt provide insight in how to transfer theoretical results from the SOM area to geometric ICA.",

keywords = "Blind source separation, Geometric ICA, Independent component analysis, Self-organizing maps",

author = "Theis, {Fabian J.} and Puntonet, {Carlos G.} and Lang, {Elmar W.}",

year = "2003",

doi = "10.1109/ICIF.2003.177411",

language = "English",

isbn = "0972184449",

series = "Proceedings of the 6th International Conference on Information Fusion, FUSION 2003",

publisher = "IEEE Computer Society",

pages = "1457--1464",

booktitle = "Proceedings of the 6th International Conference on Information Fusion, FUSION 2003",

note = "6th International Conference on Information Fusion, FUSION 2003 ; Conference date: 08-07-2003 Through 11-07-2003",

}

Theis, FJ, Puntonet, CG & Lang, EW 2003, SOMICA and geometric ICA. in Proceedings of the 6th International Conference on Information Fusion, FUSION 2003., 1255380, Proceedings of the 6th International Conference on Information Fusion, FUSION 2003, vol. 2, IEEE Computer Society, pp. 1457-1464, 6th International Conference on Information Fusion, FUSION 2003, Cairns, QLD, Australia, 8/07/03. https://doi.org/10.1109/ICIF.2003.177411

SOMICA and geometric ICA. / Theis, Fabian J.; Puntonet, Carlos G.; Lang, Elmar W.
Proceedings of the 6th International Conference on Information Fusion, FUSION 2003. IEEE Computer Society, 2003. p. 1457-1464 1255380 (Proceedings of the 6th International Conference on Information Fusion, FUSION 2003; Vol. 2).

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution › peer-review

TY - GEN

T1 - SOMICA and geometric ICA

AU - Theis, Fabian J.

AU - Puntonet, Carlos G.

AU - Lang, Elmar W.

PY - 2003

Y1 - 2003

N2 - Geometric independent component analysis (ICA) uses a weight update rule that is very similar to the self-organizing map (SOM) learning rule in the case of a trivial neighborhood function. In this paper we use this fact and present a new geometric ICA algorithm that uses a SOM for learning. The separation quality is better in comparison to other geometric algorithms, but the computational cost is higher. Furthermore, this new algorithm wilt provide insight in how to transfer theoretical results from the SOM area to geometric ICA.

AB - Geometric independent component analysis (ICA) uses a weight update rule that is very similar to the self-organizing map (SOM) learning rule in the case of a trivial neighborhood function. In this paper we use this fact and present a new geometric ICA algorithm that uses a SOM for learning. The separation quality is better in comparison to other geometric algorithms, but the computational cost is higher. Furthermore, this new algorithm wilt provide insight in how to transfer theoretical results from the SOM area to geometric ICA.

KW - Blind source separation

KW - Geometric ICA

KW - Independent component analysis

KW - Self-organizing maps

UR - http://www.scopus.com/inward/record.url?scp=84901655955&partnerID=8YFLogxK

U2 - 10.1109/ICIF.2003.177411

DO - 10.1109/ICIF.2003.177411

M3 - Conference contribution

AN - SCOPUS:84901655955

SN - 0972184449

SN - 9780972184441

T3 - Proceedings of the 6th International Conference on Information Fusion, FUSION 2003

SP - 1457

EP - 1464

BT - Proceedings of the 6th International Conference on Information Fusion, FUSION 2003

PB - IEEE Computer Society

T2 - 6th International Conference on Information Fusion, FUSION 2003

Y2 - 8 July 2003 through 11 July 2003

ER -

SOMICA and geometric ICA

Abstract

Publication series

Conference

Keywords

Access to Document

Other files and links

Fingerprint

Cite this