Efficiency of D-optimal designs for quasi-likelihood estimation in Poisson regression model with random effects

Year 2018, Volume: 47 Issue: 2, 463 - 470, 01.04.2018

S. Mehr Mansour Mehrdad Niaparast

Abstract

Optimum experimental designs are most commonly used to obtain maximum likelihood estimators of parameters. However, obtaining an explicit form of these estimators is not feasible for generalized linear mixed models (GLMMs). Hence as an alternative method to handle this issue, the quasi-likelihood method is applied to Poisson regression models with random effects, a special case of GLMMs. In this paper, we consider this model and compare D-optimal designs for quasi-likelihood estimation and maximum likelihood estimation of fixed effects parameters. The empirical results in a simulated environment suggest that the optimal designs for quasi-likelihood estimation are efficient.

Keywords

Optimal Design, Poisson Regression, Quasi-Likelihood, relative efficiency, Laplace approximation

References

• Davidian, M. and Giltinan D. M.. Nonlinear Models for Repeated Measurement Data, (London: Chapman and Hall, 1995).
• McCullagh, P., Quasi-likelihood function, The Annals of Statistics 11, 59-67, 1983.
• McCullagh, P.and Nelder, J. A., Generalized Linear Models, 2nd edn., ( London: Chapman and Hall, 1989).
• McCulloch, C. E.and Searle, S. R., Generalized Linear and Mixed Models, (New York: Wiley, 2001).
• Mielke, T., Approximations of the Fisher Information for the construction of Ecient Experimental Designs in Nonlinear Mixed Eects Models, (Ph.D. Thesis, Otto-von-Guericke University, Germany, 2012).
• Niaparast, M., On optimal design for a Poisson regression model with random intercept, Statistics and Probability Letters, textbf79, 741-747, 2009.
• Niaparast, M. and Schwabe, R., Optimal design for quasi-likelihood estimation in Poisson regression with random coecients, Journal of Statistical Planning and Inference, 143, 296- 306, 2013.
• Pinheiro, J. C. and Bates D. M., Mixed-Eects Models in S and S-Plus, (New York: Springer, 2000).
• Pukelsheim, F. (1993). Optimal design of Experiments, (New York: Wiley, 1993).
• Stufken, J. and Yang, M., On locally Optimal designs for Generalized Linear Models with group eects, The statistica sinica, 22, 1756-1786, 2012.
• Waterhouse, T. H. (2005). Optimal Experimental Design for Nonlinear and Generalised Linear Models, (PhD thesis, University of Queensland, Australia, 2005).