|
Activity Number:
|
451
|
|
Type:
|
Invited
|
|
Date/Time:
|
Wednesday, August 5, 2009 : 10:30 AM to 12:20 PM
|
|
Sponsor:
|
Noether Award Committee
|
| Abstract - #303143 |
|
Title:
|
Semiparametric Efficient Estimation in the Case-Cohort Study
|
|
Author(s):
|
Donglin Zeng*+
|
|
Companies:
|
The University of North Carolina at Chapel Hill
|
|
Address:
|
Department of Biostatistics, Chapel Hill, NC, 27514,
|
|
Keywords:
|
Case cohort ; linear transformation model ; semiparametric efficiency ; EM algorithm ; local likelihood
|
|
Abstract:
|
In the case-cohort study, some important but expensive risk covariates for failure time are only measured in a subsample randomly selected from the full cohort. Existing methods are not semiparametrically efficient when the expensive covariates depend on other continuous confounding variables. In this work, we consider estimating the risk factors in general transformation models. We propose an efficient approach by maximizing a modified likelihood function via the expectation-maximization algorithm. Particularly, the conditional densities among the covariates is obtained by maximizing a local likelihood function in the M-step. We derive the asymptotic results for the obtained estimators and show that they are consistent and asymptotically efficient. The small-sample performance of the proposed method is illustrated via numerical studies and applications to real data.
|
- 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 |
