|
Activity Number:
|
156
|
|
Type:
|
Topic Contributed
|
|
Date/Time:
|
Monday, August 1, 2016 : 10:30 AM to 12:20 PM
|
|
Sponsor:
|
Section on Nonparametric Statistics
|
|
Abstract #319723
|
View Presentation
|
|
Title:
|
Nonparametric Identifiability of Finite Mixture Model and It's Application in Alzheimer's Disease
|
|
Author(s):
|
Zheyu Wang* and Xiao-Hua Andrew Zhou
|
|
Companies:
|
The Johns Hopkins University and University of Washington
|
|
Keywords:
|
Finite mixture model ;
Identifiability ;
Latent variables
|
|
Abstract:
|
Finite mixture model have long been utilized to summarize constructs that are represented by multiple variables and are difficult to measure. It also had significant application and development in biomarker evaluation where ascertaining the true disease status can be too costly, too invasive, or too late for any treatment intervention to be effective and therefore unavailable at the time of assessment. This talk will discuss how to adopt a finite mixture model in such situation that allows for flexible component distribution as well as covariate effects that may affect component distribution or the mixing proportion. The talk will also introduce a permutation identifiability concept to discuss the nonparametric global identifiability for this type of model in general.
|
Authors who are presenting talks have a * after their name.
