|
Activity Number:
|
184
- Contributed Poster Presentations: Korean International Statistical Society
|
|
Type:
|
Contributed
|
|
Date/Time:
|
Monday, July 29, 2019 : 10:30 AM to 12:20 PM
|
|
Sponsor:
|
Korean International Statistical Society
|
|
Abstract #301722
|
|
|
Title:
|
Estimation of Semiparametric Hidden Markov Model and Multiple Testing Under Dependenct Strcuture
|
|
Author(s):
|
Joungyoun Kim* and Jong Soo Lee and Johan Lim
|
|
Companies:
|
Chungbuk National University and Department of Mathematics, University of Massachusetts at Lowell and Seoul National University
|
|
Keywords:
|
Bayesian;
variable selection;
heredity principle;
shotgun stochastic search;
strong heredity
|
|
Abstract:
|
Hidden Markov random field model is useful to take account of the dependency of multple testing hypotheses. In this paper, we propose to use the semi-parametric hidden Markov model (semi-HMM) for simultaneous hypothesis testing. We discuss its non-identifiability as in semi-parametric mixture model for the independent samples and propose a modified expectation-maximization (EM) procedure to resolve the difficulty from the non-identifiability of the model. We numerically investigate the performance of the proposed procedure in the estimation of the model and alsol compare its various multiple testing errors to two recent existing methods. In addition, we apply our procedure to analyzing two real data examples, the mass spectral experiments to differentiate the origin of a herbal medicine and the epidemic surveillance of an influenza-like illness.
|
Authors who are presenting talks have a * after their name.
