|
Activity Number:
|
189
|
|
Type:
|
Invited
|
|
Date/Time:
|
Monday, August 3, 2009 : 2:00 PM to 3:50 PM
|
|
Sponsor:
|
Section on Nonparametric Statistics
|
| Abstract - #302925 |
|
Title:
|
Functional Data Analysis Using Dirichlet Processes
|
|
Author(s):
|
Alan E. Gelfand*+
|
|
Companies:
|
Duke University
|
|
Address:
|
Department of Statistical Science, Durham, NC, 27708-0251,
|
|
Keywords:
|
clustering ; Gaussian process ; hybrid Dirichlet process ; labeling measures
|
|
Abstract:
|
Recently, Dirichlet processes have been proposed to study the problem of nonparametric estimation and clustering for a collection of functions. This talk will focus on the use of new Dirichlet label processes in conjunction with local function selection and "species" priors to address this problem. These processes will be developed formally and their properties discussed. We will analyze data on progesterone levels during menstrual cycles and brain activity levels from MRI images.
|
