|
Activity Number:
|
276
|
|
Type:
|
Topic Contributed
|
|
Date/Time:
|
Tuesday, July 31, 2007 : 10:30 AM to 12:20 PM
|
|
Sponsor:
|
IMS
|
| Abstract - #310254 |
|
Title:
|
Model Selection Using Hellinger Distance: Methods and Applications
|
|
Author(s):
|
Xiaofan Cao*+ and Haonan Wang and Hariharan Iyer
|
|
Companies:
|
Colorado State University and Colorado State University and Colorado State University
|
|
Address:
|
Campus Delivery 1877, Fort Collins, CO, 80523-1877,
|
|
Keywords:
|
Model Selection ; Hellinger Distance ; False Discovery Rate (FDR)
|
|
Abstract:
|
A model selection strategy based on an estimator of the expected Hellinger distance between an approximating model and the unknown true model is developed. Convergence and invariance properties of the proposed estimator are studied. The performance of the proposed model selection strategy is assessed in mixture distribution problems and in factorial ANOVA model selection problems using a statistical simulation study. The proposed model selection method and minimum Hellinger distance estimation are applied to the problem of estimation of the false discovery rate (FDR) in Microarray data analysis. The methods are illustrated using real data examples. The R codes for their implementation will be made available upon request.
|
