|
Activity Number:
|
121
|
|
Type:
|
Contributed
|
|
Date/Time:
|
Monday, August 3, 2009 : 8:30 AM to 10:20 AM
|
|
Sponsor:
|
IMS
|
| Abstract - #304466 |
|
Title:
|
Finite Mixture of Heteroscedastic Single Index Models
|
|
Author(s):
|
Peng Zeng*+
|
|
Companies:
|
Auburn University
|
|
Address:
|
Department of Mathematics and Statistics, Auburn, AL, 36849,
|
|
Keywords:
|
finite mixture model ; single index model ; semi-parametric regression ; EM algorithm ; clustering analysis
|
|
Abstract:
|
Modeling high dimensional data is challenging when it is difficult to specify an appropriate parametric model due to the lack of enough prior knowledge. In this talk, we consider a semiparametric model for regression. Assume that the whole population consists of several subpopulations, and each subpopulation only depends on the predictors via its one linear combination. Each subpopulation is modeled by a single index model with heteroscedasticity, and thus the whole population is modeled by a finite mixture of heteroscedastic single index models. The model is fitted via a variant of EM algorithm. Some theoretical results and implementation concerns are discussed. Both simulation studies and real examples are used to demonstrate the application of this model.
|
- 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 |
