|
Activity Number:
|
291
|
|
Type:
|
Contributed
|
|
Date/Time:
|
Tuesday, August 8, 2006 : 10:30 AM to 12:20 PM
|
|
Sponsor:
|
Section on Nonparametric Statistics
|
| Abstract - #306225 |
|
Title:
|
Model Selection in Accelerated Failure Time Models with Nonlinear Covariate Effects
|
|
Author(s):
|
Chenlei Leng*+ and Shuangge Ma
|
|
Companies:
|
National University of Singapore and University of Washington
|
|
Address:
|
6 Science Drive 2, Singapore, 117546, Singapore
|
|
Keywords:
|
accelerated failure time ; COSSO ; Stute's estimator
|
|
Abstract:
|
As an alternative to the Cox model, the accelerated failure time (AFT) model assumes that the event time of interest depends on the covariates through a regression function. We investigate the AFT model with nonparametric covariate effects, when model selection is desirable. Formulated in the framework of smoothing spline ANOVA, the COSSO method with the Stute estimate can achieve a sparse representation of functional decomposition, by utilizing a reproducing kernel Hilbert norm penalty. Computational algorithms and theoretical properties of the method are investigated. The usefulness of the methodology is demonstrated via simulation studies and a real clinical data set.
|
