Abstract
This paper is concerned with the problem of pseudo-Bayesian D-optimal designs for the first-order Poisson mixed model for longitudinal data with time-dependent correlated errors. A standard approximate covariance matrix of the parameter estimation is obtained based on the quasi-likelihood method. Furthermore, to overcome the dependence of pseudo-Bayesian D-optimal designs on the choice of the prior mean, a hierarchical pseudo-Bayesian D-optimal designs based on the hierarchical prior distribution of unknown parameters is proposed. The results show that the optimal number of time points depends on both the interclass autoregressive coefficients and different cost constraints. The relative efficiency of equidistant designs compared with the hierarchical pseudo-Bayesian D-optimal designs is also discussed.
Similar content being viewed by others
References
Abebe HT, Tan FES, Breukelen VGJP, Berger MPF (2014) Bayesian D-optimal designs for the two parameter logistic mixed effects model. Comput Stat Data Anal 71(1):1066–1076
Atkinson AC, Donev AN, Tobias RD (2007) Optimal experimental designs, with SAS. Oxford University Press, New York
Casella G, Berger RL (2001) Statistical inference, 2nd edn. Duxbury Press, London
Chi EM, Reinsel GC (1989) Models for longitudinal data with random effects and AR(1) errors. J Am Stat Assoc 84(406):452–459
Fang K-T, Wang Y (1994) Number-theoretic methods in statistics. Chapman and Hall, London
Fedorov VV, Leonov SL (2014) Optimal design for nonlinear response models. Chapman and Hall/CRC Press, Boca Raton
Ford I, Torsney B, Wu CFJ (1992) The use of a canonical form in the construction of locally optimal designs for non-linear problems. J R Stat Soc Ser B 54(2):569–583
Khuri AI, Mukherjee B, Sinha BK, Ghosh M (2006) Design issues for generalized linear models: a review. Stat Sci 21(3):376–399
Li W, Nachtsheim CJ (2000) Model-robust factorial designs. Technometrics 42:345–352
Liang KY, Zeger SL (1986) Longitudinal data analysis using generalized linear models. Biometrika 78:13–22
Lin TI, Wang WL (2013) Multivariate skew-normal at linear mixed models for multi-outcome longitudinal data. Stat Model 13(3):199–221
McCullagh P, Nelder JA (1989) Generalized linear models, 2nd edn. Chapman and Hall, New York
Mcgree JM, Eccleston JA (2012) Robust designs for poisson regression models. Technometrics 54(1):64–72
Mcgree JM, Duffull SB, Eccleston JA, Ward LC (2007) Optimal designs for studying bioimpedance. Physiol Meas 28(12):1465
Molenberghs G, Verbeke G (2005) Models for discrete longitudinal data. Springer, New York
Niaparast M (2010) Optimal designs for mixed effects poisson regression models. Ph.D. thesis, Magdeburg University, Germany
Niaparast M, Schwabe R (2013) Optimal design for quasi-likelihood estimation in poisson regression with random coefficients. J Stat Plan Inference 143:296–306
Ouwens MJNM, Tan FES, Berger MPF (2006) A maximin criterion for the logistic random intercept model with covariates. J Stat Plan Inference 136(3):962–981
Roy A (2006) Estimating correlation coefficient between two variables with repeated observations using mixed effects model. Biom J 48(2):286–301
Russell KG, Woods DC, Lewis SM, Eccleston JA (2009) D-optimal designs for poisson regression models. Stat Sin 19(2):2831–2845
Ryan EG, Drovandi CC, Pettitt AN (2015) Simulation-based fully bayesian experimental design for mixed effects models. Comput Stat Data Anal 92:26–39
Tan FES, Berger MPF (1999) Optimal allocation of time points for the random effects model. Commun Stat Simul Comput 28(2):517–540
Tekle FB (2008) D-optimal designs for prospective cohort studies. Ph.D. thesis, Department of Methodology and Statistics, University of Maastricht, The Netherlands
Tekle FB, Tan FES, Berger MPF (2008) Maximin D-optimal designs for binary longitudinal responses. Comput Stat Data Anal 52(12):5253–5262
Wang W-L, Fan T-H (2010) ECM-based maximum likelihood inference for multivariate linear mixed models with autoregressive errors. Comput Stat Data Anal 54(5):1328–1341
Wang Y, Myers RH, Smoth EP, Ye K (2006) D-optimal designs for poisson regression models. J Stat Plan Inference 136:2831–2845
Wedderburn RWM (1974) Quasi-likelihood functions, generalized linear models, and the gauss-newton method. Biometrika 61(3):439–447
Author information
Authors and Affiliations
Corresponding author
Additional information
This work was supported by NSFC Grants 11471216 and 11871143.
Rights and permissions
About this article
Cite this article
Jiang, HY., Yue, RX. Pseudo-Bayesian D-optimal designs for longitudinal Poisson mixed models with correlated errors. Comput Stat 34, 71–87 (2019). https://doi.org/10.1007/s00180-018-0834-7
Received:
Accepted:
Published:
Issue Date:
DOI: https://doi.org/10.1007/s00180-018-0834-7