UUM Repository | Universiti Utara Malaysian Institutional Repository
FAQs | Feedback | Search Tips | Sitemap

A simulation study of a parametric mixture model of three different distributions to analyze heterogeneous survival data


Mohammed, Yusuf Abbakar and Yatim, Bidin and Ismail, Suzilah (2013) A simulation study of a parametric mixture model of three different distributions to analyze heterogeneous survival data. Modern Applied Science, 7 (7). pp. 1-9. ISSN 1913-1844

[img]
Preview
PDF
Available under License Creative Commons Attribution.

Download (257kB) | Preview

Abstract

In this paper a simulation study of a parametric mixture model of three different distributions is considered to model heterogeneous survival data.Some properties of the proposed parametric mixture of Exponential, Gamma and Weibull are investigated.The Expectation Maximization Algorithm (EM) is implemented to estimate the maximum likelihood estimators of three different postulated parametric mixture model parameters.The simulations are performed by simulating data sampled from a population of three component parametric mixture of three different distributions, and the simulations are repeated 10, 30, 50, 100 and 500 times to investigate the consistency and stability of the EM scheme.The EM Algorithm scheme developed is able to estimate the parameters of the mixture which are very close to the parameters of the postulated model.The repetitions of the simulation give parameters closer and closer to the postulated models, as the number of repetitions increases, with relatively small standard errors.

Item Type: Article
Uncontrolled Keywords: survival time analysis, maximum likelihood, em-algorithm, mixture model, simulation, exponential distribution, gamma distribution, weibull distribution
Subjects: Q Science > QA Mathematics
Divisions: College of Arts and Sciences
Depositing User: Prof. Madya Dr. Bidin Yatim
Date Deposited: 22 Apr 2014 01:37
Last Modified: 22 Apr 2014 01:37
URI: http://repo.uum.edu.my/id/eprint/10772

Actions (login required)

View Item View Item