|
Activity Number:
|
499
|
|
Type:
|
Topic Contributed
|
|
Date/Time:
|
Thursday, August 2, 2007 : 8:30 AM to 10:20 AM
|
|
Sponsor:
|
Section on Nonparametric Statistics
|
| Abstract - #308849 |
|
Title:
|
Minimum Distance Estimator of Judgment-Ranking Probabilities in a Ranked-Set Sample
|
|
Author(s):
|
Omer Ozturk*+
|
|
Companies:
|
The Ohio State University
|
|
Address:
|
Department of Statistics, Columbus , OH, 43210,
|
|
Keywords:
|
imperfect ranking ; doubly stochastic ; best linear unbiased estimator ; calibration ; doubly stochastic matrix ; perfect ranking
|
|
Abstract:
|
In ranked set sampling procedure, ranking error is almost an unavoidable reality. Thus, a reasonable statistical inference should address the validity of the procedure under imperfect ranking. There are several models that describe the ranking mechanism. One of these models is introduced by Bohn and Wolfe(1994) and later extended by Frey(2006). Their model constructs the judgment class distributions as a mixture distributions of the actual order statistics. In this talk, we first develop a minimum distance estimator for the mixing parameters of this mixture distribution. Then we use this estimated model to calibrate the estimator and tests to reduce effects of ranking errors on statistical procedures.
|
