A New Hybrid Estimation Method for the Generalized Exponential Distribution
Date
2013-07-10
Authors
Journal Title
Journal ISSN
Volume Title
Publisher
Abstract
The generalized exponential distribution (GED) is a popular distribution for analyzing lifetime data and has been introduced and studied quite extensively. It can be used as an alternative to gamma or Weibull distribution in many situations. As one of the most important research areas of GED, the estimation of the parameters has been discussed by many authors. Among the existing methods, the maximum likelihood method and Bayesian method are the methods that many authors recommend to use since they can provide balanced and good performances for different sample sizes and parameter values.
In order to improve the estimation in terms of bias and mean squared error (MSE), and to simplify the computation for the Bayesian method, in this thesis, we introduce a new hybrid estimation method for the GED, which is based on the idea of minimizing a goodness-of- fit measure and incorporating useful maximum likelihood information. Through simulation, we compare the new hybrid method with the MLE method and the Bayesian methods as well as other existing methods, and show that this new hybrid method can not only reduce the estimation bias but also improve the MSE especially for the small sample case.
Description
Keywords
Statistics
Citation
Zhu, S. (2013). A New Hybrid Estimation Method for the Generalized Exponential Distribution (Master's thesis, University of Calgary, Calgary, Canada). Retrieved from https://prism.ucalgary.ca. doi:10.11575/PRISM/24866