|
Activity Number:
|
470
|
|
Type:
|
Contributed
|
|
Date/Time:
|
Wednesday, August 9, 2006 : 2:00 PM to 3:50 PM
|
|
Sponsor:
|
Biometrics Section
|
| Abstract - #306184 |
|
Title:
|
Analysis of Linear Transformation Models with Covariate Transformations
|
|
Author(s):
|
Chunpeng Fan*+ and Jason P. Fine
|
|
Companies:
|
University of Wisconsin-Madison and University of Wisconsin-Madison
|
|
Address:
|
Department of Statistics and Department of Biostatistics and Medical Informatics, Madison, WI, 53706,
|
|
Keywords:
|
linear transformation model ; covariate transformation ; nuisance parameter under alternative ; semiparametric model
|
|
Abstract:
|
The linear transformation model is a natural extension of the linear regression model to permit nonparametric response transformation. There has been much work on such models without covariate transformations. However, in many applications, both response and covariate transformation may be required. We propose inferences of the regression and transformation parameters for the linear transformation model with unknown parametric covariate transformations. Since transformation parameters vanish under the null of no covariate effect, such tests are nonstandard. When covariate effects are non-zero, the estimated regression and transformation parameters are consistent and asymptotically normal. Simulation studies show that the tests and estimators perform well with realistic sample size. An application to the well known GAGurine data shows improved goodness of fit compared to earlier analyses.
|
