The Binomial Parameter Estimation by Using Weighted Method of Two Bayesian Estimators

Article Sidebar

PDF
Keywords:
Parameter estimation Binomial distribution Bayesian estimator

Main Article Content

Wararit Panichkitkosolkul
Department of Mathematics and Statistics, Faculty of Science and Technology, Thammasat University, Pathum Thani, 12121, Thailand.

Abstract

This paper proposes a new binomial parameter estimation by using the Weighted
method of two Bayesian estimators: Bayesian estimator given Beta(1/2,1/2) prior distribution
and Bayesian estimator given Beta(2,2) prior distribution. The used weighted value has the
minimum variance of the estimator. In addition, this research compares three parameter
estimation methods. These methods are: Bayesian method given Beta(1/2,1/2) prior
distribution (B1), Bayesian method given Beta(2,2) prior distribution (B2), and Weighted
method of two Bayesian estimators (W). Monte Carlo simulation is used to investigate the
behavior of this new binomial parameter estimation based on mean square errors (MSE).
Simulation results are as follows: For all sample sizes, the MSE of B1 method is the lowest
when p \fn_phv \leq 0.20 and the MSE of B2 method is the lowest when 0.30 \fn_phv < p \fn_phv < 0.50 . For the
sample sizes at least 50, The MSE of W method is the lowest when p is equal to 0.25.

Article Details

How to Cite
Panichkitkosolkul, W. (2015). The Binomial Parameter Estimation by Using Weighted Method of Two Bayesian Estimators. Science & Technology Asia, 14(2), 35–41. Retrieved from https://ph02.tci-thaijo.org/index.php/SciTechAsia/article/view/41365
Issue
Vol.14 No.2 (April-June 2009)
Section
Articles