|
Activity Number:
|
232
|
|
Type:
|
Contributed
|
|
Date/Time:
|
Monday, August 3, 2009 : 2:00 PM to 3:50 PM
|
|
Sponsor:
|
WNAR
|
| Abstract - #303385 |
|
Title:
|
Detecting Subclusters in Outliers by Cluster Analysis
|
|
Author(s):
|
Dongseok Choi*+ and Zhixin Kang and Carrie Nielson and Eric Orwoll and George Tiao
|
|
Companies:
|
Oregon Health & Science University and The University of North Carolina at Pembroke and Oregon Health & Science University and Oregon Health & Science University and The University of Chicago
|
|
Address:
|
3181 SW Sam Jack Park Road, Portland, OR, 97239,
|
|
Keywords:
|
clustering ; outliers ; subcluster ; mixture distribution ; Normal distribution ; likelihood ratio test
|
|
Abstract:
|
New clustering methods have been proposed in recent years, mainly motivated by microarray data. This new generation of clustering methods has two distinct properties: 1) does not require predefined number of groups and 2) finds small tight groups while not forcing isolated points into a group. We tested these new and traditional clustering methods whether a clustering method can identify one main group, i.e. non-outliers, and small clusters of outliers. Simulation results will be presented and an interactive demonstration of real data will be given.
|
