Variable Selection for Model-Based High-Dimensional Clustering and Its Application to Microarray Data.

by: Sijian Wang, Ji Zhu

Biometrics (26 October 2007)

View FullText article

Blackwell Synergy, DOI, Pubmed, Hubmed, Wiley InterScience

Reviews [Write a review of this article]

There are no reviews of this article

Find related articles from these CiteULike users

jyuh, amargoli

Find related articles with these CiteULike tags

dimensionality-reduction, microarray, statistics

Abstract

Variable selection in high-dimensional clustering analysis is an important yet challenging problem. In this article, we propose two methods that simultaneously separate data points into similar clusters and select informative variables that contribute to the clustering. Our methods are in the framework of penalized model-based clustering. Unlike the classical L(1)-norm penalization, the penalty terms that we propose make use of the fact that parameters belonging to one variable should be treated as a natural "group." Numerical results indicate that the two new methods tend to remove noninformative variables more effectively and provide better clustering results than the L(1)-norm approach.

@article{citeulike:2087524, abstract = {Variable selection in high-dimensional clustering analysis is an important yet challenging problem. In this article, we propose two methods that simultaneously separate data points into similar clusters and select informative variables that contribute to the clustering. Our methods are in the framework of penalized model-based clustering. Unlike the classical L(1)-norm penalization, the penalty terms that we propose make use of the fact that parameters belonging to one variable should be treated as a natural "group." Numerical results indicate that the two new methods tend to remove noninformative variables more effectively and provide better clustering results than the L(1)-norm approach.}, address = {Department of Biostatistics, University of Michigan, Ann Arbor, Michigan 48109, U.S.A.}, author = {Wang, Sijian and Zhu, Ji }, citeulike-article-id = {2087524}, doi = {http://dx.doi.org/10.1111/j.1541-0420.2007.00922.x}, issn = {0006-341X}, journal = {Biometrics}, keywords = {dimensionality-reduction}, month = {October}, posted-at = {2008-05-14 06:48:18}, priority = {2}, title = {Variable Selection for Model-Based High-Dimensional Clustering and Its Application to Microarray Data.}, url = {http://dx.doi.org/10.1111/j.1541-0420.2007.00922.x}, year = {2007} }

RIS record

TY - JOUR ID - citeulike:2087524 L3 - citeulike-article-id:2087524 TI - Variable Selection for Model-Based High-Dimensional Clustering and Its Application to Microarray Data. JF - Biometrics AD - Department of Biostatistics, University of Michigan, Ann Arbor, Michigan 48109, U.S.A. SN - 0006-341X N2 - Variable selection in high-dimensional clustering analysis is an important yet challenging problem. In this article, we propose two methods that simultaneously separate data points into similar clusters and select informative variables that contribute to the clustering. Our methods are in the framework of penalized model-based clustering. Unlike the classical L(1)-norm penalization, the penalty terms that we propose make use of the fact that parameters belonging to one variable should be treated as a natural "group." Numerical results indicate that the two new methods tend to remove noninformative variables more effectively and provide better clustering results than the L(1)-norm approach. KW - dimensionality-reduction AU - Wang, Sijian AU - Zhu, Ji PY - 2007/10/26/ UR - http://dx.doi.org/10.1111/j.1541-0420.2007.00922.x ER -

RIS BibTeX