|
Activity Number:
|
297
|
|
Type:
|
Invited
|
|
Date/Time:
|
Tuesday, August 4, 2009 : 10:30 AM to 12:20 PM
|
|
Sponsor:
|
SSC
|
| Abstract - #302894 |
|
Title:
|
Composite Likelihood EM Algorithm in High-Dimensional Data Analysis
|
|
Author(s):
|
Peter X.K. Song*+
|
|
Companies:
|
University of Michigan
|
|
Address:
|
Department of Biostatistics, Ann Arbor, ME, 48109-2029,
|
|
Keywords:
|
gene network ; hidden Markov model ; latent variable ; missing data ; time-course microarray data
|
|
Abstract:
|
The method of composite likelihood is useful to deal with estimation and inference in parametric models with high-dimensional data, where the full likelihood approach is greatly challenged by intractable computational complexity. We develop an extension of the EM algorithm in the framework of composite likelihood estimation in the presence of missing data or latent variables. We establish three key theoretical properties of the composite likelihood EM (CLEM) algorithm, including the ascent property, the algorithmic convergence and the convergence rate. Two applications of the proposed algorithm are presented for illustration, including high-dimensional copula models with incomplete data and multivariate hidden Markov models with microarray data. Both analytic and empirical performances of the proposed CLEM algorithm are demonstrated simulation studies and real world data examples.
|
- 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 2009 program |
