|
Activity Number:
|
522
|
|
Type:
|
Contributed
|
|
Date/Time:
|
Wednesday, August 5, 2009 : 2:00 PM to 3:50 PM
|
|
Sponsor:
|
Section on Statistical Computing
|
| Abstract - #304388 |
|
Title:
|
Semiparametric Efficient Estimation and EM Algorithm for Partially Linear Model with Missing Data
|
|
Author(s):
|
Mingyu Li*+ and Minge Xie
|
|
Companies:
|
Rutgers University and Rutgers University
|
|
Address:
|
35 Hawthorne Dr Apt A2, Somerset, NJ, 08873,
|
|
Keywords:
|
Semiparametric efficiency ; Estimating equation ; Smoothing
|
|
Abstract:
|
This article studies a partially linear regression model with missing response variable and develops efficient semiparametric inference for the parametric component of the model. The missingness considered here includes a broad range of missing patterns. We propose an algorithm which runs iteratively between fitting parametric components and fitting nonparametric components while holding the others fixed. We develop an EM algorithm to estimate the parametric part by a semiparametric estimating equation and estimate the nonparametric part by smoothing methods. The asymptotic distribution theory for the estimator of parametric part is obtained and it is shown that our estimator is asymptotically normally distributed. Furthermore, we prove that the asymptotic covariance of the estimator achieves the semiparametric lower bound. The methodology is illustrated by numerical examples.
|
