|
Activity Number:
|
586
|
|
Type:
|
Topic Contributed
|
|
Date/Time:
|
Thursday, August 6, 2009 : 10:30 AM to 12:20 PM
|
|
Sponsor:
|
International Chinese Statistical Association
|
| Abstract - #303860 |
|
Title:
|
Pitman Closeness of Estimators and Predictors for Exponential Distribution
|
|
Author(s):
|
Katherine Davies*+ and Narayanaswamy Balakrishnan and Jerome P. Keating
|
|
Companies:
|
University of Manitoba and McMaster University and The University of Texas at San Antonio
|
|
Address:
|
Department of Statistics, Winnipeg, MB, R3T 2N2, Canada
|
|
Keywords:
|
order statistics ; Pitman closeness ; estimator ; predictor
|
|
Abstract:
|
Comparisons of best linear unbiased have been carried out over the last fifty years since the ground breaking work of Lloyd in 1952. These comparisons have been made under many different criteria across different families of distributions. Herein, we add to the literature for the comparison of different linear forms of the order statistics from a Type-II censored sample under the Pitman closeness (nearness) criterion. We focus special interest on the exponential distribution and establish that the best linear invariant estimator (BLIE) is Pitman closeness inadmissible to the best linear unbiased estimator (BLUE). Moreover, based on a Type-II censored sample, we compare the best linear unbiased predictor (BLUP) of future order statistics with the best linear invariant predictor (BLIP). Under the Pitman closeness comparison the predictor of choice is quite frequently the BLIP.
|
