Aarhus University Seal / Aarhus Universitets segl

Multiple Clustering Views via Constrained Projections

Research output: Contribution to book/anthology/report/proceedingArticle in proceedingsResearchpeer-review

Standard

Multiple Clustering Views via Constrained Projections. / Dang, Xuan-Hong; Assent, Ira; Bailey, James.

2012 SIAM International Conference on Data Mining: 3rd MultiClust Workshop: Discovering, Summarizing and Using Multiple Clusterings. ed. / Emmanuel Müller; Thomas Seidl; Arthur Zimek. Society for Industrial and Applied Mathematics, 2012. p. 23-30.

Research output: Contribution to book/anthology/report/proceedingArticle in proceedingsResearchpeer-review

Harvard

Dang, X-H, Assent, I & Bailey, J 2012, Multiple Clustering Views via Constrained Projections. in E Müller, T Seidl & A Zimek (eds), 2012 SIAM International Conference on Data Mining: 3rd MultiClust Workshop: Discovering, Summarizing and Using Multiple Clusterings. Society for Industrial and Applied Mathematics, pp. 23-30, MultiClust Workshop, Anaheim, United States, 28/04/2012. <http://siam.omnibooksonline.com/2012datamining/data/papers/WS06.pdf>

APA

Dang, X-H., Assent, I., & Bailey, J. (2012). Multiple Clustering Views via Constrained Projections. In E. Müller, T. Seidl, & A. Zimek (Eds.), 2012 SIAM International Conference on Data Mining: 3rd MultiClust Workshop: Discovering, Summarizing and Using Multiple Clusterings (pp. 23-30). Society for Industrial and Applied Mathematics. http://siam.omnibooksonline.com/2012datamining/data/papers/WS06.pdf

CBE

Dang X-H, Assent I, Bailey J. 2012. Multiple Clustering Views via Constrained Projections. Müller E, Seidl T, Zimek A, editors. In 2012 SIAM International Conference on Data Mining: 3rd MultiClust Workshop: Discovering, Summarizing and Using Multiple Clusterings. Society for Industrial and Applied Mathematics. pp. 23-30.

MLA

Dang, Xuan-Hong, Ira Assent and James Bailey "Multiple Clustering Views via Constrained Projections"., Müller, Emmanuel Seidl, Thomas Zimek, Arthur (editors). 2012 SIAM International Conference on Data Mining: 3rd MultiClust Workshop: Discovering, Summarizing and Using Multiple Clusterings. Society for Industrial and Applied Mathematics. 2012, 23-30.

Vancouver

Dang X-H, Assent I, Bailey J. Multiple Clustering Views via Constrained Projections. In Müller E, Seidl T, Zimek A, editors, 2012 SIAM International Conference on Data Mining: 3rd MultiClust Workshop: Discovering, Summarizing and Using Multiple Clusterings. Society for Industrial and Applied Mathematics. 2012. p. 23-30

Author

Dang, Xuan-Hong ; Assent, Ira ; Bailey, James. / Multiple Clustering Views via Constrained Projections. 2012 SIAM International Conference on Data Mining: 3rd MultiClust Workshop: Discovering, Summarizing and Using Multiple Clusterings. editor / Emmanuel Müller ; Thomas Seidl ; Arthur Zimek. Society for Industrial and Applied Mathematics, 2012. pp. 23-30

Bibtex

@inproceedings{60314d0cee2a4f8bb92b37b2711e61ae,
title = "Multiple Clustering Views via Constrained Projections",
abstract = "Clustering, the grouping of data based on mutual similarity,is often used as one of principal tools to analyze and understanddata. Unfortunately, most conventional techniquesaim at finding only a single clustering over the data. Formany practical applications, especially those being describedin high dimensional data, it is common to see that the datacan be grouped into different yet meaningful ways. Thisgives rise to the recently emerging research area of discoveringalternative clusterings. In this preliminary work, wepropose a novel framework to generate multiple clusteringviews. The framework relies on a constrained data projectionapproach by which we ensure that a novel alternativeclustering being found is not only qualitatively strong butalso distinctively different from a reference clustering solution.We demonstrate the potential of the proposed frameworkusing both synthetic and real world datasets and discusssome future research directions with the approach.",
author = "Xuan-Hong Dang and Ira Assent and James Bailey",
year = "2012",
language = "English",
pages = "23--30",
editor = "M{\"u}ller, {Emmanuel } and Seidl, {Thomas } and Zimek, {Arthur }",
booktitle = "2012 SIAM International Conference on Data Mining",
publisher = "Society for Industrial and Applied Mathematics",
note = "MultiClust Workshop ; Conference date: 28-04-2012 Through 28-04-2012",

}

RIS

TY - GEN

T1 - Multiple Clustering Views via Constrained Projections

AU - Dang, Xuan-Hong

AU - Assent, Ira

AU - Bailey, James

N1 - Conference code: 3

PY - 2012

Y1 - 2012

N2 - Clustering, the grouping of data based on mutual similarity,is often used as one of principal tools to analyze and understanddata. Unfortunately, most conventional techniquesaim at finding only a single clustering over the data. Formany practical applications, especially those being describedin high dimensional data, it is common to see that the datacan be grouped into different yet meaningful ways. Thisgives rise to the recently emerging research area of discoveringalternative clusterings. In this preliminary work, wepropose a novel framework to generate multiple clusteringviews. The framework relies on a constrained data projectionapproach by which we ensure that a novel alternativeclustering being found is not only qualitatively strong butalso distinctively different from a reference clustering solution.We demonstrate the potential of the proposed frameworkusing both synthetic and real world datasets and discusssome future research directions with the approach.

AB - Clustering, the grouping of data based on mutual similarity,is often used as one of principal tools to analyze and understanddata. Unfortunately, most conventional techniquesaim at finding only a single clustering over the data. Formany practical applications, especially those being describedin high dimensional data, it is common to see that the datacan be grouped into different yet meaningful ways. Thisgives rise to the recently emerging research area of discoveringalternative clusterings. In this preliminary work, wepropose a novel framework to generate multiple clusteringviews. The framework relies on a constrained data projectionapproach by which we ensure that a novel alternativeclustering being found is not only qualitatively strong butalso distinctively different from a reference clustering solution.We demonstrate the potential of the proposed frameworkusing both synthetic and real world datasets and discusssome future research directions with the approach.

M3 - Article in proceedings

SP - 23

EP - 30

BT - 2012 SIAM International Conference on Data Mining

A2 - Müller, Emmanuel

A2 - Seidl, Thomas

A2 - Zimek, Arthur

PB - Society for Industrial and Applied Mathematics

T2 - MultiClust Workshop

Y2 - 28 April 2012 through 28 April 2012

ER -