Comparison Studies of Estimators for the Generalized Gamma Distribution and New Findings
J. Chang (TU Delft - Electrical Engineering, Mathematics and Computer Science)
Piao Chen – Mentor (TU Delft - Statistics)
More Info
expand_more
Other than for strictly personal use, it is not permitted to download, forward or distribute the text or part of it, without the consent of the author(s) and/or copyright holder(s), unless the work is under an open content license such as Creative Commons.
Abstract
The Generalized Gamma Distribution (GGD) is a three-parameter distribution with desirable properties. For certain values of the parameters, the GGD can reduce to the gamma, exponential and lognormal distribution, among others. This makes it a flexible distribution that can be used in various scenarios. The problem with implementation of the GGD is that parameter estimation through maximum likelihood gives inaccurate results. In this thesis I analyse four proposed parameter estimation methods and compare them in different scenarios. Lawless' proposed estimation method produced the most accurate solutions, but other estimation methods may be more favorable depending on necessity for faster runtime, ease of implementation or available sample sizes. An attempt at devising a new parameter estimation method is made, but ultimately failed.