|
Activity Number:
|
30
|
|
Type:
|
Contributed
|
|
Date/Time:
|
Sunday, August 2, 2009 : 2:00 PM to 3:50 PM
|
|
Sponsor:
|
Section on Nonparametric Statistics
|
| Abstract - #304592 |
|
Title:
|
An Efficient Estimator for an Additive Quantile Regression Model
|
|
Author(s):
|
Dawit Zerom*+
|
|
Companies:
|
California State University, Fullerton
|
|
Address:
|
P.O. Box 92834-6848, Fullerton, CA, 92834,
|
|
Keywords:
|
Additive Quantile ; Kernel method ; Oracle efficient
|
|
Abstract:
|
We introduce a two-step kernel estimator for estimating additive components of a nonparametric additive quantile regression model when the data are drawn from a strictly stationary absolutely regular process. We provide the asymptotic distribution the proposed estimator and show the estimator is oracle efficient in the sense that each additive component has the same asymptotic bias and variance as in the case when all other components were known. On a practical side, the proposed estimator is also computationally convenient requiring two easy steps. Simulations confirm that the two-step estimator is superior to existing kernel estimators of the additive conditional quantile model. We also provide a demonstrative empirical application to modeling the conditional quantile of travel times of emergency medical services.
|
