|
Activity Number:
|
380
|
|
Type:
|
Contributed
|
|
Date/Time:
|
Wednesday, August 9, 2006 : 8:30 AM to 10:20 AM
|
|
Sponsor:
|
ENAR
|
| Abstract - #305922 |
|
Title:
|
Efficient Estimation in Semiparametric Generalized Linear Model for Longitudinal Data
|
|
Author(s):
|
Lu Wang*+ and Xihong Lin and Andrea Rotnitzky
|
|
Companies:
|
Harvard University and Harvard School of Public Health and Harvard University
|
|
Address:
|
655 Huntington Ave., Sph 2, 4th floor, Boston, MA, 02115,
|
|
Keywords:
|
semiparametric generalized linear model ; longitudinal/clustered data ; generalized estimating equation ; iterative kernel GEE estimator ; semiparametric efficient score ; semiparametric information bound
|
|
Abstract:
|
We consider estimation in semiparametric generalized linear model for longitudinal or clustered data. Conventional profile-kernel method fails to yield a semiparametric efficient estimator for the coefficients of parametric covariates. Wang, Carroll and Lin proposed an iterative kernel generalized estimating equation (GEE) estimator, which accounted for within-cluster correlation and is more efficient. We derive the semiparametric efficient score function and the semiparametric information bound in general scenarios under the semiparametric conditional mean model. We discuss the asymptotic properties of the iterative profile-kernel GEE estimator and show that it is semiparametric efficient. Simulations will be performed to demonstrate our results. Further extensions to the missing data case will be considered.
|
