Your Cart

Your cart is empty

Home  >  Volume 50

7. COM-POISSON MIXED EFFECT MODEL FOR OVER-DISPERSED CORRELATED COUNT DATA: A SIMULATION STUDY by H.G. Dikko, D.A. Shobanke, O.E. Asiribo, and B.B. Alhaji. Volume 50 (March, 2019 Issue)
Sale price: $0.00
Quantity:

COM-POISSON MIXED EFFECT MODEL FOR OVER-DISPERSED CORRELATED COUNT DATA: A SIMULATION STUDY

H.G. Dikko, D.A. Shobanke, O.E. Asiribo, and B.B. Alhaji.

Department of Statistics, Ahmadu Bello University, Zaria Nigeria.

Department of Mathematical Sciences, Federal University Lokoja, Kogi State.

Department of Statistics, University of Agriculture, Abeokuta Nigeria.

4Nigerian Defence Academy, Kaduna Nigeria.

Abstract

The Poisson regression is popularly used to model count data. However practically, count data do not always satisfy the assumption of equality of mean and variance which is an important property of the Poisson distribution. The Poisson-Gamma regression and the Conway-Maxwell Poisson (COM-Poisson) regression are some of the proposed remedies for handling under and over dispersed independent data. The COM-Poisson regression has been recently extended for effective modeling of correlated count data regardless of the level of dispersion by using the COM-Poisson as the base distribution in a Generalized linear mixed effects model (GLMM). In this paper, the COM-Poisson GLMM alongside Poisson and Negative-Binomial GLMMs are evaluated using simulation studies. Error of estimation is used as evaluation criteria. The simulation results show that our model produced results that are similar and in small sample setting better than some of the existing methods which imply that our model can be used as a credible alternative.

Keywords: Over-dispersion, Correlated data, Poisson regression, Mixed-effect, Simulation study

click here to download full abstract
Close
Loading...