Far East Journal of Theoretical Statistics
Volume 21, Issue 2, Pages 255 - 266
(March 2007)
|
|
ON LINEAR REGRESSION UNDER THE PARTIAL KOZIOL-GREEN MODEL OF
RANDOM CENSORSHIP
Ke Wu (U. S. A.)
|
Abstract: In survival analysis, a model of informative censoring is the
Koziol-Green (KG) model, where the survival function of the censoring times is
some power of the survival function of the lifetimes. A generalization of the KG
model is the partial Koziol-Green (PKG) model which allows the lifetimes to be
censored by two types of variables, one of which censors in an informative way
and the other one in a non-informative way. In this paper we give a new method
of estimating parameters in the linear regression model under the PKG model. The
partial ACL estimator for the survival function under the PKG model is used in
the proposed method, rather than the Kaplan-Meier [10] estimator used for the
general random censoring model in the Buckley and James’ [3] procedure for
estimating parameters in linear regression with censored data. Simulations
comparing the two methods are described. |
Keywords and phrases: linear regression, partial Koziol-Green model, random
censorship, partial ACL estimator, product limit estimator. |
|
Number of Downloads: 250 | Number of Views: 587 |
|