Article ID | Journal | Published Year | Pages | File Type |
---|---|---|---|---|
10525215 | Journal of Statistical Planning and Inference | 2005 | 16 Pages |
Abstract
This is a study of the behaviors of the naive bootstrap and the Bayesian bootstrap clones designed to approximate the sampling distribution of the Aalen-Johansen estimator of a non-homogeneous censored Markov chain. The study shows that the approximations based on the Bayesian bootstrap clones and the naive bootstrap are first-order asymptotically equivalent. The two bootstrap methods are illustrated by a marketing example, and their performance is validated by a Monte Carlo experiment.
Related Topics
Physical Sciences and Engineering
Mathematics
Applied Mathematics
Authors
Cheng-Der Fuh, Edward H. Ip,