|
Activity Number:
|
604
|
|
Type:
|
Contributed
|
|
Date/Time:
|
Thursday, August 6, 2009 : 10:30 AM to 12:20 PM
|
|
Sponsor:
|
Section on Physical and Engineering Sciences
|
| Abstract - #305675 |
|
Title:
|
Optimal Goodness-of-Fit Tests for Recurrent Event Data
|
|
Author(s):
|
Russell Stocker*+
|
|
Companies:
|
Mississippi State University
|
|
Address:
|
Department of Mathematics and Statistics, Mississippi, MS, 39762,
|
|
Keywords:
|
recurrent events ; reliability ; goodness of fit ; counting processes
|
|
Abstract:
|
In this talk we consider a study of lifetime data whose units experience multiple ordered events of the same nature. These events are commonly referred to as recurrent events. A class of models used to analyze recurrent event data is presented. We propose goodness-of-fit tests for the null hypothesis that the baseline intensity function belongs to a parametric family of intensities indexed by a parameter. These tests are based on a general class of weighted empirical processes. The asymptotic properties of these processes are given. Goodness-of-fits are then constructed based on these processes. Optimality results with regards to local power are presented for these tests. A real data set is used to illustrate the proposed tests.
|
