Skip to main content

Likelihood-based and Bayesian methods for Tweedie compound Poisson linear mixed models

Abstract

The Tweedie compound Poisson distribution is a subclass of the exponential dispersion family with a power variance function, in which the value of the power index lies in the interval (1,2). It is well known that the Tweedie compound Poisson density function is not analytically tractable, and numerical procedures that allow the density to be accurately and fast evaluated did not appear until fairly recently. Unsurprisingly, there has been little statistical literature devoted to full maximum likelihood inference for Tweedie compound Poisson mixed models. To date, the focus has been on estimation methods in the quasi-likelihood framework. Further, Tweedie compound Poisson mixed models involve an unknown variance function, which has a significant impact on hypothesis tests and predictive uncertainty measures. The estimation of the unknown variance function is thus of independent interest in many applications. However, quasi-likelihood-based methods are not well suited to this task. This paper presents several likelihood-based inferential methods for the Tweedie compound Poisson mixed model that enable estimation of the variance function from the data. These algorithms include the likelihood approximation method, in which both the integral over the random effects and the compound Poisson density function are evaluated numerically; and the latent variable approach, in which maximum likelihood estimation is carried out via the Monte Carlo EM algorithm, without the need for approximating the density function. In addition, we derive the corresponding Markov Chain Monte Carlo algorithm for a Bayesian formulation of the mixed model. We demonstrate the use of the various methods through a numerical example, and conduct an array of simulation studies to evaluate the statistical properties of the proposed estimators.

This is a preview of subscription content, access via your institution.

Fig. 1
Fig. 2
Fig. 3
Fig. 4
Fig. 5

References

  1. Barndorff-Nielsen, O.: Information and Exponential Families in Statistical Theory. Wiley, Chichester (1978)

    MATH  Google Scholar 

  2. Bates, D., Mächler, M., Bolker, B.: Fitting linear mixed-effects models using lme4. J. Stat. Softw. (2012, forthcoming)

  3. Booth, J.G., Hobert, J.P.: Maximizing generalized linear mixed model likelihoods with an automated Monte Carlo EM algorithm. J. R. Stat. Soc. B 61, 265–285 (1999)

    Article  MATH  Google Scholar 

  4. Breslow, N.E., Clayton, D.G.: Approximate inference in generalized linear mixed models. J. Am. Stat. Assoc. 88(421), 9–25 (1993)

    MATH  Google Scholar 

  5. Browne, W.J., Draper, D.: A comparison of Bayesian and likelihood-based methods for fitting multilevel models. Bayesian Anal. 1(3), 473–514 (2006)

    MathSciNet  Google Scholar 

  6. Cox, D.R., Reid, N.: Parameter orthogonality and approximate conditional inference. J. R. Stat. Soc. B 49, 1–39 (1987)

    MathSciNet  MATH  Google Scholar 

  7. Davidian, M., Carroll, R.J.: Variance function estimation. J. Am. Stat. Assoc. 82, 1079–1091 (1987)

    MathSciNet  Article  MATH  Google Scholar 

  8. de Silva, H.N., Hall, A.J., Tustin, D.S., Gandar, P.W.: Analysis of distribution of root length density of apple trees on different dwarfing rootstocks. Ann. Bot. 83, 335–345 (1999)

    Article  Google Scholar 

  9. Dunn, P.K., Smyth, G.K.: Series evaluation of Tweedie exponential dispersion models densities. Stat. Comput. 15, 267–280 (2005)

    MathSciNet  Article  Google Scholar 

  10. Dunn, P.K., Smyth, G.K.: Evaluation of Tweedie exponential dispersion model densities by Fourier inversion. Stat. Comput. 18, 73–86 (2008)

    MathSciNet  Article  Google Scholar 

  11. Gelman, A., Rubin, D.B.: Inference from iterative simulation using multiple sequences (with discussion). Stat. Sci. 7, 457–511 (1992)

    Article  Google Scholar 

  12. Gelman, A., Carlin, J.B., Stern, H.S., Rubin, D.B.: Bayesian Data Analysis. Chapman and Hall, London (2003)

    Google Scholar 

  13. Jørgensen, B.: Exponential dispersion models (with discussion). J. R. Stat. Soc. B 49, 127–162 (1987)

    Google Scholar 

  14. Jørgensen, B., Knudsen, S.J.: Parameter orthogonality and bias adjustment for estimating functions. Scand. J. Stat. 31, 93–114 (2004)

    Article  Google Scholar 

  15. Laird, N.M., Ware, J.H.: Random-effects models for longitudinal data. Biometrics 38, 963–974 (1982)

    Article  MATH  Google Scholar 

  16. Levine, R.A., Casella, G.: Implementations of the Monte Carlo EM algorithm. J. Comput. Graph. Stat. 10(3), 422–439 (2001)

    MathSciNet  Article  Google Scholar 

  17. Liao, J.G., Lipsitz, S.R.: A type of restricted maximum likelihood estimator of variance components in generalised linear mixed models. Biometrika 89(2), 401–409 (2002)

    MathSciNet  Article  MATH  Google Scholar 

  18. Liu, Q., Pierce, D.A.: A note on Gauss-Hermite quadrature. Biometrika 81(3), 624–629 (1994)

    MathSciNet  MATH  Google Scholar 

  19. McCullagh, P., Nelder, J.: Generalized Linear Models. Chapman and Hall, Boca Raton (1989)

    MATH  Book  Google Scholar 

  20. McCullagh, P., Tibshirani, R.: A simple method for the adjustment of profile likelihoods. J. R. Stat. Soc. B 52, 325–344 (1990)

    MathSciNet  MATH  Google Scholar 

  21. McCulloch, C.E.: Maximum likelihood algorithms for generalized linear mixed models. J. Am. Stat. Assoc. 92, 162–170 (1997)

    MathSciNet  Article  MATH  Google Scholar 

  22. McCulloch, C.E., Searle, S.R.: Generalized, Linear and Mixed Models. Wiley, New York (2001)

    MATH  Google Scholar 

  23. Meng, X.-L., Rubin, D.B.: Maximum likelihood estimation via the ECM algorithm: A general framework. Biometrika 80(2), 267–278 (1993)

    MathSciNet  Article  MATH  Google Scholar 

  24. Nelder, J.A., Pregibon, D.: An extended quasi-likelihood function. Biometrika 74, 221–232 (1987)

    MathSciNet  Article  MATH  Google Scholar 

  25. Patterson, H.D., Thompson, R.N.: Recovery of inter-block information when block sizes are unequal. Biometrika 58, 545–554 (1971)

    MathSciNet  Article  MATH  Google Scholar 

  26. Peters, G.W., Shevchenko, P.V., Wüthrich, M.V.: Model uncertainty in claims reserving within Tweedie’s compound Poisson models. ASTIN Bull. 39(1), 1–33 (2009)

    MathSciNet  Article  MATH  Google Scholar 

  27. Pinheiro, J.C., Chao, E.C.: Efficient Laplacian and adaptive Gaussian quadrature algorithms for multilevel generalized linear mixed models. J. Comput. Graph. Stat. 15(1), 58–81 (2006)

    MathSciNet  Article  Google Scholar 

  28. Robert, C.P., Casella, G.: Monte Carlo Statistical Methods. Springer, New York (2004)

    Book  MATH  Google Scholar 

  29. Smyth, G.K.: Regression analysis of quantity data with exact zeros. In: Proceedings of the Second Australia-Japan Workshop on Stochastic Models in Engineering, Technology and Management, pp. 572–580. Technology Management Centre, University of Queensland (1996)

  30. Tierney, L., Kadane, J.B.: Accurate approximation for posterior moments and marginal densities. J. Am. Stat. Assoc. 81, 82–86 (1986)

    MathSciNet  Article  MATH  Google Scholar 

  31. Wei, G.C.G., Tanner, M.A.: A Monte Carlo implementation of the EM algorithm and the poor man’s data augmentation algorithms. J. Am. Stat. Assoc. 85, 699–704 (1990)

    Article  Google Scholar 

  32. Zeger, S.L., Karim, R.M.: Generalized linear models with random effects; a Gibbs sampling approach. J. Am. Stat. Assoc. 86(413), 79–86 (1991)

    MathSciNet  Article  Google Scholar 

Download references

Acknowledgements

The author thanks James Guszcza for his review of an earlier version of the paper.

Author information

Affiliations

Authors

Corresponding author

Correspondence to Yanwei Zhang.

Rights and permissions

Reprints and Permissions

About this article

Cite this article

Zhang, Y. Likelihood-based and Bayesian methods for Tweedie compound Poisson linear mixed models. Stat Comput 23, 743–757 (2013). https://doi.org/10.1007/s11222-012-9343-7

Download citation

Keywords

  • Adaptive Gauss-Hermite quadrature
  • Extended quasi-likelihood
  • Laplace approximation
  • Monte Carlo EM
  • Maximum likelihood estimation
  • Mixed models
  • Penalized quasi-likelihood
  • Tweedie compound Poisson distribution