Activity Number:
|
101
|
Type:
|
Topic Contributed
|
Date/Time:
|
Monday, August 12, 2002 : 10:30 AM to 12:20 PM
|
Sponsor:
|
IMS
|
Abstract - #300874 |
Title:
|
Model Selection in Semiparametric Regression Analysis
|
Author(s):
|
Runze Li*+ and Jianqing Fan
|
Affiliation(s):
|
Pennsylvania State University and Chinese University, Hong Kong/UNC-CH
|
Address:
|
326 Thomas Building, University Park, Pennsylvania, PA 16802-2111, USA
|
Keywords:
|
LASSO ; MODEL SELECTION ; PENALIZED LEAST SQUARES ; SCAD
|
Abstract:
|
In this talk, I will propose a class of variable selection procedures for parametric portion in the semiparametric regression models. The proposed procedures are distinguished from other existing ones in that they select significant variables and estimate unknown parameters simultaneously. Sampling properties of the resulting estimate are established. A robust standard error formula is derived via using sandwich formula and empirically tested. Statistical inference for nonparametric portion in the models will be briefly discussed. Empirical performances of the proposed inference procedures are investigated via extensive Monte Carlo simulations.
|