The views expressed here are those of the individual authors and not necessarily those of the JSM sponsors, their officers, or their staff.
Abstract Details
Activity Number:
|
118
|
Type:
|
Topic Contributed
|
Date/Time:
|
Monday, August 1, 2011 : 8:30 AM to 10:20 AM
|
Sponsor:
|
IMS
|
Abstract - #302389 |
Title:
|
A New Initialization Procedure for the EM Algorithm in Gaussian Mixture Models
|
Author(s):
|
Volodymyr Melnykov*+ and Igor Melnykov
|
Companies:
|
North Dakota State University and Colorado State University at Pueblo
|
Address:
|
Department of Statistics, Fargo, ND, 58108-6050, USA
|
Keywords:
|
EM algorithm ;
Gaussian mixture model ;
initialization
|
Abstract:
|
The success of convergence of the EM algorithm in finite mixture models depends on effective initialization. There are multiple approaches proposed in literature that deal with this problem. However, there is no method that can be preferred over the others in all cases. We propose a new procedure for Gaussian mixtures that can be seen as a generalization of popular emEM and Rnd-EM algorithms. The suggested method demonstrates promising performance and good results in many cases. We illustrate the proposed approach on several simulated and classification datasets.
|
The address information is for the authors that have a + after their name.
Authors who are presenting talks have a * after their name.
Back to the full JSM 2011 program
|
2011 JSM Online Program Home
For information, contact jsm@amstat.org or phone (888) 231-3473.
If you have questions about the Continuing Education program, please contact the Education Department.