|
Activity Number:
|
244
|
|
Type:
|
Contributed
|
|
Date/Time:
|
Tuesday, July 31, 2007 : 8:30 AM to 10:20 AM
|
|
Sponsor:
|
Biometrics Section
|
| Abstract - #308400 |
|
Title:
|
Empirical Bayes Estimators in Stratified Random Sampling/Ranked Set Sampling
|
|
Author(s):
|
Tetsuji Ohyama*+ and Jimmy A. Doi and Takashi Yanagawa
|
|
Companies:
|
Kurume University and California Polytechnic State University and Kurume University
|
|
Address:
|
67 Asahimachi, Kurume City, International, 830-0011, Japan
|
|
Keywords:
|
biased estimator ; infinite population ; U statistics ; relative efficiency
|
|
Abstract:
|
In this paper, we discuss the estimation of population characteristics in a stratified random sampling in an infinite population framework and consider a use of prior information by empirical Bayes method. Underlying distribution is assumed unknown, and U-statistics is employed for constructing the estimator incorporated with the prior values of the characteristics. The optimal empirical Bayes estimator is obtained, but it contains unknown parameters and those parameters must be replaced in practice. Simulation is conducted to show the gains in efficiency of the proposed estimator, showing the gain is 1.1~1.5 times larger than the unbiased estimator in terms of the relative efficiency if prior values are close to the true value.
|
