|
Activity Number:
|
147
|
|
Type:
|
Contributed
|
|
Date/Time:
|
Monday, August 7, 2006 : 10:30 AM to 12:20 PM
|
|
Sponsor:
|
Section on Physical and Engineering Sciences
|
| Abstract - #305636 |
|
Title:
|
Robust Prediction and Extrapolation Designs for Censored Data
|
|
Author(s):
|
Xiaojian Xu*+
|
|
Companies:
|
University of Alberta
|
|
Address:
|
109 Lees Ave., Apt. 619, Ottawa, ON, K1S 5L5, Canada
|
|
Keywords:
|
regression design ; extrapolation ; censoring ; nonsmooth optimization ; accelerated life testing ; maximum likelihood estimation
|
|
Abstract:
|
We present the construction of robust designs for a possibly misspecified generalized linear regression model when the data are censored. The minimax designs are found for maximum likelihood estimation in the context of both prediction and extrapolation problems with or without restraint of design unbiasedness. This paper extends preceding work of robust designs for complete data by incorporating censoring and maximum likelihood estimation. It also broadens former work of robust designs for censored data by considering both nonlinearity and much more arbitrary uncertainty in fitted regression response and by dropping all restrictions on the structure of regressors. Solutions are derived by a nonsmooth optimization technique. A typical example in accelerated life testing also is demonstrated.
|
