Estimation of exponential Pareto parameters

Document Type : Research Paper

Authors

1 Middle Technical University, Institute of Administration Al Russafa, Baghdad, Iraq

2 Department of Statistics, Administration and Economics College, Kerbala University, Iraq

3 University of Misan, College of Administration and Economic, Misan, Iraq

Abstract

In this paper, we simulated the methods of estimation of Pareto exponential distribution parameters, using various experiments including default parameter values (0.5, 1, 5), sample sizes (10,250,100), and Iterations (D = 1000). The Maximum likelihood method and the standard Bayes method were used with informative prior distribution (Gamma distribution) and the non-information prior distribution (Uniform distribution), and symmetric loss function  (quadratic loss function), and asymmetric (General Entropy loss function). The Bayes method gave complex mathematical formulas that were solved by Lindley approximation. The results of the simulation experiments showed the advantage of the standard Bayes method in the information prior and the quadratic loss function.

Keywords