On the Performance of Dirichlet Prior Mixture of Generalized Linear Mixed Models for Zero Truncated Count Data

Authors

O. S. Adesina, Department of Mathematics and Statistics, Redeemer’s University, Osun State, NigeriaFollow
K. S. Adekeye, Department of Mathematics and Statistics, Redeemer’s University, Osun State, NigeriaFollow
A. F. Adedotun, Department of Mathematics, Covenant University, Ota, Ogun State, NigeriaFollow
N. O. Adeboye, Department of Statistics, Osun State University, Osogbo, Osun State, NigeriaFollow
P. O. Ogundile, Department of Mathematics, Covenant University, Ota, Ogun State, NigeriaFollow
O. A. Odetunmibi, Department of Mathematics, Covenant University, Ota, Ogun State, NigeriaFollow

Author Country (or Countries)

Nigeria

Abstract

In this study, the performance of Dirichlet Process Mixture of Generalized Linear Mixed Models (DPMGLMMs) was examined against some competing models for fitting zero-truncated count data. The Bayesian models such as Monte Carlo Markov Chain GLMMs, Bayesian Discrete Weibull and the frequentists models such as Zero truncated Poisson, Zero truncated Binomial and Zero truncated Geometric models were compared with the proposed DPMGLMMs model. Simulation and life count data from health domain was used to compare the performance of DPMGLMM with the Bayesian and frequentist models considered in this study. The results showed that the DPMGLMM outperformed other models considered for fitting count data that is truncated at zero.

Digital Object Identifier (DOI)

http://dx.doi.org/10.18576/jsap/120324

Recommended Citation

S. Adesina, O.; S. Adekeye, K.; F. Adedotun, A.; O. Adeboye, N.; O. Ogundile, P.; and A. Odetunmibi, O. (2023) "On the Performance of Dirichlet Prior Mixture of Generalized Linear Mixed Models for Zero Truncated Count Data," Journal of Statistics Applications & Probability: Vol. 12: Iss. 3, Article 24.
DOI: http://dx.doi.org/10.18576/jsap/120324
Available at: https://digitalcommons.aaru.edu.jo/jsap/vol12/iss3/24