Feature weighted rival penalized EM for Gaussian mixture clustering: Automatic feature and model selections in a single paradigm

Yiu Ming Cheung; Hong Zeng

doi:10.1109/ICCIAS.2006.294213

Feature weighted rival penalized EM for Gaussian mixture clustering: Automatic feature and model selections in a single paradigm

Yiu Ming Cheung^*, Hong Zeng

^*Corresponding author for this work

Department of Computer Science

Research output: Chapter in book/report/conference proceeding › Conference proceeding › peer-review

7 Citations (Scopus)

Abstract

Recently, the Rival Penalized Expectation-Maximization (RPEM) algorithm has demonstrated its powerful capability to perform the model selection automatically in the context of mixture model. However, the performance may be degraded when irrelevant variables are included. To overcome this drawback, we adopt the concept of feature salience as the feature weight to measure the relevance to the clusters in the subspace, and integrate it into the RPEM algorithm. The proposed algorithm distinguishes the probably redundant features and estimate the number of clusters automatically and simultaneously in a single learning paradigm. Experiments conducted on both synthetic and benchmark real data set have shown the efficacy of the proposed algorithm.

Original language	English
Title of host publication	2006 International Conference on Computational Intelligence and Security, ICCIAS 2006
Publisher	IEEE Computer Society
Pages	633-638
Number of pages	6
ISBN (Print)	1424406056, 9781424406050
DOIs	https://doi.org/10.1109/ICCIAS.2006.294213
Publication status	Published - 3 Nov 2006
Event	2006 International Conference on Computational Intelligence and Security, CIS 2006 - Guangzhou, China Duration: 3 Oct 2006 → 6 Oct 2006 https://ieeexplore.ieee.org/xpl/conhome/4072023/proceeding https://link.springer.com/book/10.1007/978-3-540-74377-4

Publication series

Name	2006 International Conference on Computational Intelligence and Security, ICCIAS 2006
Volume	1

Conference

Conference	2006 International Conference on Computational Intelligence and Security, CIS 2006
Country/Territory	China
City	Guangzhou
Period	3/10/06 → 6/10/06
Internet address	https://ieeexplore.ieee.org/xpl/conhome/4072023/proceeding https://link.springer.com/book/10.1007/978-3-540-74377-4

Scopus Subject Areas

Computer Science(all)
Control and Systems Engineering

Access to Document

10.1109/ICCIAS.2006.294213

Cite this

Cheung, Y. M., & Zeng, H. (2006). Feature weighted rival penalized EM for Gaussian mixture clustering: Automatic feature and model selections in a single paradigm. In 2006 International Conference on Computational Intelligence and Security, ICCIAS 2006 (pp. 633-638). Article 4072166 (2006 International Conference on Computational Intelligence and Security, ICCIAS 2006; Vol. 1). IEEE Computer Society. https://doi.org/10.1109/ICCIAS.2006.294213

Cheung, Yiu Ming ; Zeng, Hong. / Feature weighted rival penalized EM for Gaussian mixture clustering : Automatic feature and model selections in a single paradigm. 2006 International Conference on Computational Intelligence and Security, ICCIAS 2006. IEEE Computer Society, 2006. pp. 633-638 (2006 International Conference on Computational Intelligence and Security, ICCIAS 2006).

@inproceedings{5367d7c5d8034720aedb0a6f07379f5c,

title = "Feature weighted rival penalized EM for Gaussian mixture clustering: Automatic feature and model selections in a single paradigm",

abstract = "Recently, the Rival Penalized Expectation-Maximization (RPEM) algorithm has demonstrated its powerful capability to perform the model selection automatically in the context of mixture model. However, the performance may be degraded when irrelevant variables are included. To overcome this drawback, we adopt the concept of feature salience as the feature weight to measure the relevance to the clusters in the subspace, and integrate it into the RPEM algorithm. The proposed algorithm distinguishes the probably redundant features and estimate the number of clusters automatically and simultaneously in a single learning paradigm. Experiments conducted on both synthetic and benchmark real data set have shown the efficacy of the proposed algorithm.",

author = "Cheung, {Yiu Ming} and Hong Zeng",

year = "2006",

month = nov,

day = "3",

doi = "10.1109/ICCIAS.2006.294213",

language = "English",

isbn = "1424406056",

series = "2006 International Conference on Computational Intelligence and Security, ICCIAS 2006",

publisher = "IEEE Computer Society",

pages = "633--638",

booktitle = "2006 International Conference on Computational Intelligence and Security, ICCIAS 2006",

address = "United States",

note = "2006 International Conference on Computational Intelligence and Security, CIS 2006 ; Conference date: 03-10-2006 Through 06-10-2006",

url = "https://ieeexplore.ieee.org/xpl/conhome/4072023/proceeding, https://link.springer.com/book/10.1007/978-3-540-74377-4",

}

Cheung, YM & Zeng, H 2006, Feature weighted rival penalized EM for Gaussian mixture clustering: Automatic feature and model selections in a single paradigm. in 2006 International Conference on Computational Intelligence and Security, ICCIAS 2006., 4072166, 2006 International Conference on Computational Intelligence and Security, ICCIAS 2006, vol. 1, IEEE Computer Society, pp. 633-638, 2006 International Conference on Computational Intelligence and Security, CIS 2006, Guangzhou, China, 3/10/06. https://doi.org/10.1109/ICCIAS.2006.294213

Feature weighted rival penalized EM for Gaussian mixture clustering: Automatic feature and model selections in a single paradigm. / Cheung, Yiu Ming; Zeng, Hong.
2006 International Conference on Computational Intelligence and Security, ICCIAS 2006. IEEE Computer Society, 2006. p. 633-638 4072166 (2006 International Conference on Computational Intelligence and Security, ICCIAS 2006; Vol. 1).

Research output: Chapter in book/report/conference proceeding › Conference proceeding › peer-review

TY - GEN

T1 - Feature weighted rival penalized EM for Gaussian mixture clustering

T2 - 2006 International Conference on Computational Intelligence and Security, CIS 2006

AU - Cheung, Yiu Ming

AU - Zeng, Hong

PY - 2006/11/3

Y1 - 2006/11/3

N2 - Recently, the Rival Penalized Expectation-Maximization (RPEM) algorithm has demonstrated its powerful capability to perform the model selection automatically in the context of mixture model. However, the performance may be degraded when irrelevant variables are included. To overcome this drawback, we adopt the concept of feature salience as the feature weight to measure the relevance to the clusters in the subspace, and integrate it into the RPEM algorithm. The proposed algorithm distinguishes the probably redundant features and estimate the number of clusters automatically and simultaneously in a single learning paradigm. Experiments conducted on both synthetic and benchmark real data set have shown the efficacy of the proposed algorithm.

AB - Recently, the Rival Penalized Expectation-Maximization (RPEM) algorithm has demonstrated its powerful capability to perform the model selection automatically in the context of mixture model. However, the performance may be degraded when irrelevant variables are included. To overcome this drawback, we adopt the concept of feature salience as the feature weight to measure the relevance to the clusters in the subspace, and integrate it into the RPEM algorithm. The proposed algorithm distinguishes the probably redundant features and estimate the number of clusters automatically and simultaneously in a single learning paradigm. Experiments conducted on both synthetic and benchmark real data set have shown the efficacy of the proposed algorithm.

UR - http://www.scopus.com/inward/record.url?scp=38549175107&partnerID=8YFLogxK

U2 - 10.1109/ICCIAS.2006.294213

DO - 10.1109/ICCIAS.2006.294213

M3 - Conference proceeding

AN - SCOPUS:38549175107

SN - 1424406056

SN - 9781424406050

T3 - 2006 International Conference on Computational Intelligence and Security, ICCIAS 2006

SP - 633

EP - 638

BT - 2006 International Conference on Computational Intelligence and Security, ICCIAS 2006

PB - IEEE Computer Society

Y2 - 3 October 2006 through 6 October 2006

ER -

Cheung YM, Zeng H. Feature weighted rival penalized EM for Gaussian mixture clustering: Automatic feature and model selections in a single paradigm. In 2006 International Conference on Computational Intelligence and Security, ICCIAS 2006. IEEE Computer Society. 2006. p. 633-638. 4072166. (2006 International Conference on Computational Intelligence and Security, ICCIAS 2006). doi: 10.1109/ICCIAS.2006.294213

Feature weighted rival penalized EM for Gaussian mixture clustering: Automatic feature and model selections in a single paradigm

Abstract

Publication series

Conference

Scopus Subject Areas

Access to Document

Other files and links

Fingerprint

Cite this