|
Activity Number:
|
119
|
|
Type:
|
Contributed
|
|
Date/Time:
|
Monday, August 7, 2006 : 8:30 AM to 10:20 AM
|
|
Sponsor:
|
Section on Nonparametric Statistics
|
| Abstract - #306360 |
|
Title:
|
Semiparametric Mixture Approach for the Measurement Error Problem in the Presence of Additional Error-Free Covariate
|
|
Author(s):
|
Byungtae Seo*+ and Bruce G. Lindsay
|
|
Companies:
|
The Pennsylvania State University and The Pennsylvania State University
|
|
Address:
|
326 Thomas Building, University Park, PA, 16802,
|
|
Keywords:
|
measurement error ; semiparametric ; mixture ; kernel estimator ; Kullback-Leibler ; MLE*
|
|
Abstract:
|
In usual statistical problems, our interest is to find functional relationship between response variable Y and covariate X. Measurement error problem occurs when measuring true covariate X is expensive or impossible to directly measure. In this case, sometimes another predictor W is available such that W represents the true predictor X but with some measurement error. We studied this measurement error problem when there are two types of covariates, one is measured with error and another is measured without error. It is known that semi-parametric method produces inconsistent estimators. The new method proposed based on kernel smoothing. The consistency of the estimators proposed is obtained under fairly mild conditions. A small simulation shows that usual semi-parametric mixture approach produces inconsistent estimators and suggests that the proposed estimators perform well.
|
