A comparative performance analysis of self organizing maps on weight initializations using different strategies

Haripriya Harikumar; R Devisree; Dinesh Pooja; Prema Nedungadi

doi:10.1109/ICACC.2015.75

A comparative performance analysis of self organizing maps on weight initializations using different strategies

Haripriya Harikumar, R Devisree, Dinesh Pooja, Prema Nedungadi

Machine Learning and Robotics

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

Abstract

Self Organizing Maps perform clustering of data based on unsupervised learning. It is of concern that initialization of the weight vector contributes significantly to the performance of SOM and since real world datasets being high-dimensional, the complexity of SOM tend to increase tremendously leading to increased time consumption as well. Our work focuses on the analysis of different weight initialization strategies and various dimensionality reduction measures with the intent to make SOM flexible for handling high-dimensional datasets. We use two methods of comparison, one on projected space and another before projection. The datasets used are real world datasets taken from UCI repository.

Original language	English
Title of host publication	Proceedings 2015 Fifth International Conference on Advances in Computing and Communications
Subtitle of host publication	ICACC 2015
Place of Publication	Los Alamitos, CA
Publisher	IEEE
Pages	434-438
Number of pages	5
ISBN (Electronic)	9781467369930
ISBN (Print)	9781467369947
DOIs	https://doi.org/10.1109/ICACC.2015.75
Publication status	Published - 17 Mar 2016
Event	5th International Conference on Advances in Computing and Communications - Kochi, India Duration: 2 Sept 2015 → 4 Sept 2015

Conference

Conference	5th International Conference on Advances in Computing and Communications
Abbreviated title	ICACC
Country/Territory	India
City	Kochi
Period	2/09/15 → 4/09/15

Keywords

unsupervised learning
SOM
PCA
NLPCA
FCM
weight initialization

Access to Document

10.1109/ICACC.2015.75

Cite this

@inproceedings{201fbb151a414023844a96977aae0575,

title = "A comparative performance analysis of self organizing maps on weight initializations using different strategies",

abstract = "Self Organizing Maps perform clustering of data based on unsupervised learning. It is of concern that initialization of the weight vector contributes significantly to the performance of SOM and since real world datasets being high-dimensional, the complexity of SOM tend to increase tremendously leading to increased time consumption as well. Our work focuses on the analysis of different weight initialization strategies and various dimensionality reduction measures with the intent to make SOM flexible for handling high-dimensional datasets. We use two methods of comparison, one on projected space and another before projection. The datasets used are real world datasets taken from UCI repository.",

keywords = "unsupervised learning, SOM, PCA, NLPCA, FCM, weight initialization",

author = "Haripriya Harikumar and R Devisree and Dinesh Pooja and Prema Nedungadi",

year = "2016",

month = mar,

day = "17",

doi = "10.1109/ICACC.2015.75",

language = "English",

isbn = "9781467369947",

pages = "434--438",

booktitle = "Proceedings 2015 Fifth International Conference on Advances in Computing and Communications",

publisher = "IEEE",

address = "United States",

note = "5th International Conference on Advances in Computing and Communications , ICACC ; Conference date: 02-09-2015 Through 04-09-2015",

}

Harikumar, H, Devisree, R, Pooja, D & Nedungadi, P 2016, A comparative performance analysis of self organizing maps on weight initializations using different strategies. in Proceedings 2015 Fifth International Conference on Advances in Computing and Communications: ICACC 2015. IEEE, Los Alamitos, CA, pp. 434-438, 5th International Conference on Advances in Computing and Communications , Kochi, India, 2/09/15. https://doi.org/10.1109/ICACC.2015.75

A comparative performance analysis of self organizing maps on weight initializations using different strategies. / Harikumar, Haripriya; Devisree, R; Pooja, Dinesh et al.
Proceedings 2015 Fifth International Conference on Advances in Computing and Communications: ICACC 2015. Los Alamitos, CA: IEEE, 2016. p. 434-438.

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

TY - GEN

T1 - A comparative performance analysis of self organizing maps on weight initializations using different strategies

AU - Harikumar, Haripriya

AU - Devisree, R

AU - Pooja, Dinesh

AU - Nedungadi, Prema

PY - 2016/3/17

Y1 - 2016/3/17

N2 - Self Organizing Maps perform clustering of data based on unsupervised learning. It is of concern that initialization of the weight vector contributes significantly to the performance of SOM and since real world datasets being high-dimensional, the complexity of SOM tend to increase tremendously leading to increased time consumption as well. Our work focuses on the analysis of different weight initialization strategies and various dimensionality reduction measures with the intent to make SOM flexible for handling high-dimensional datasets. We use two methods of comparison, one on projected space and another before projection. The datasets used are real world datasets taken from UCI repository.

AB - Self Organizing Maps perform clustering of data based on unsupervised learning. It is of concern that initialization of the weight vector contributes significantly to the performance of SOM and since real world datasets being high-dimensional, the complexity of SOM tend to increase tremendously leading to increased time consumption as well. Our work focuses on the analysis of different weight initialization strategies and various dimensionality reduction measures with the intent to make SOM flexible for handling high-dimensional datasets. We use two methods of comparison, one on projected space and another before projection. The datasets used are real world datasets taken from UCI repository.

KW - unsupervised learning

KW - SOM

KW - PCA

KW - NLPCA

KW - FCM

KW - weight initialization

U2 - 10.1109/ICACC.2015.75

DO - 10.1109/ICACC.2015.75

M3 - Conference contribution

SN - 9781467369947

SP - 434

EP - 438

BT - Proceedings 2015 Fifth International Conference on Advances in Computing and Communications

PB - IEEE

CY - Los Alamitos, CA

T2 - 5th International Conference on Advances in Computing and Communications

Y2 - 2 September 2015 through 4 September 2015

ER -

A comparative performance analysis of self organizing maps on weight initializations using different strategies

Abstract

Conference

Keywords

Access to Document

Fingerprint

Cite this