|
Activity Number:
|
112
|
|
Type:
|
Contributed
|
|
Date/Time:
|
Monday, August 3, 2009 : 8:30 AM to 10:20 AM
|
|
Sponsor:
|
Section on Nonparametric Statistics
|
| Abstract - #304736 |
|
Title:
|
Estimation and Variable Selection for Semiparametric Additive Partial Linear Models
|
|
Author(s):
|
Xiang Liu*+ and Li (Lily) Wang and Hua Liang
|
|
Companies:
|
University of Rochester Medical Center and The University of Georgia and University of Rochester
|
|
Address:
|
601 Elmwood Avenue, Box 630 , Rochester, NY, 14642,
|
|
Keywords:
|
BIC ; False Selection Rate ; LASSO ; Penalized likelihood ; SCAD ; Variable Selection
|
|
Abstract:
|
We propose an estimation method for semiparametric additive partial linear models by using polynomial splines to approximate nonparametric functions, and derive the asymptotic normality of the resulting estimators. We also develop a variable selection procedure to identify important linear components using the smoothly clipped absolute deviation penalty (SCAD). The SCAD-based estimators are shown to have oracle property. Simulations are conducted to examine the performance of our approach as compared to other variable selection methods such as BIC, LASSO and controlling false selection rate. The proposed approach is also applied to a real data from a nutritional epidemiology study. We find that log beta-carotene is nonlinearly related to intake of cholesterol and age, but linearly related to BMI, intake of fiber, gender and smoking status.
|
