|
Activity Number:
|
204
|
|
Type:
|
Contributed
|
|
Date/Time:
|
Monday, July 30, 2007 : 2:00 PM to 3:50 PM
|
|
Sponsor:
|
Section on Quality and Productivity
|
| Abstract - #310079 |
|
Title:
|
Within-Sample Prediction of Future Failure Times Based on Type-II Censored Samples from the Birnbaum-Saunders Distribution
|
|
Author(s):
|
Kevin S. McCarter*+
|
|
Companies:
|
Louisiana State University
|
|
Address:
|
161 Agricultural Administration Bldg, Baton Rouge, LA, 70803-5606,
|
|
Keywords:
|
Prediction ; Birnbaum-Saunders ; Censoring ; Simulation ; Type-II ; Calibration
|
|
Abstract:
|
In this study we investigate procedures for the within-sample prediction of the largest failure time in Type-II censored samples from the Birnbaum-Saunders distribution. The first procedure we consider uses the parameter-estimated conditional distribution of the largest order statistic to construct the prediction intervals. We show that intervals constructed in this way do not achieve nominal confidence levels, with the actual coverage probabilities depending upon several factors including the sample size and the observed number of failure times. Logistic regression is used to quantify this dependence, and the resulting model is used to develop a calibrated interval prediction procedure with improved true coverage probability. Simulation results show that the calibrated intervals attain their nominal confidence levels at most of the simulation factor settings considered in the study.
|
