Advertisement

TEST

, Volume 28, Issue 2, pp 543–564 | Cite as

A flexible class of parametric distributions for Bayesian linear mixed models

  • Mohsen Maleki
  • Darren WraithEmail author
  • Reinaldo B. Arellano-Valle
Original Paper

Abstract

In this paper, we consider a linear mixed effect model (LMM) assuming that the random effect and error terms follow an unrestricted skew-normal generalized-hyperbolic (SUNGH) distribution. The SUNGH is a broad class of flexible distributions that includes various other well-known asymmetric and symmetric families and provides a high degree of flexibility for the modeling of complex multivariate data with different directions and degrees of asymmetry, kurtosis and heavy tails. The choice of the best fitting distribution can proceed quite naturally through parameter estimation or by placing constraints on specific parameters and assessing using model choice criteria. We estimate parameters of the LMM using a Bayesian approach and examine the performance of the proposed methodology on simulated and real data from a clinical trial on treatment options for schizophrenia (Lapierre et al. Acta Psychiatric Scandinavica 82:72–76, 1990; Ho and Lin Biom J 52(4):449–469, 2010).

Keywords

Bayesian analysis Linear mixed effect model MCMC method Unrestricted skew-normal generalized-hyperbolic distribution Unrestricted skew-normal distribution 

Mathematics Subject Classification

62J02 62P10 62F15 

Notes

Acknowledgements

The authors would like to thank the AE and anonymous reviewers for their suggestions, corrections and encouragement, which helped us to improve earlier versions of the manuscript.

References

  1. Arellano-Valle RB, Azzalini A (2006) On the unification of families of skew-normal distributions. Scand J Stat 33:561–574MathSciNetCrossRefzbMATHGoogle Scholar
  2. Arellano-Valle RB, Genton MG (2005) On fundamental skew distributions. J Multivar Anal 96:93–116MathSciNetCrossRefzbMATHGoogle Scholar
  3. Arellano-Valle RB, Bolfarine H, Lachos GH (2005) Skew-normal linear mixed model. J Data Sci 3:415–438Google Scholar
  4. Arellano-Valle RB, Bolfarine H, Lachos GH (2007) Bayesian inference for skew-normal linear mixed model. J Appl Stat 33:561–574MathSciNetGoogle Scholar
  5. Azzalini A (2014) The skew-normal and related families. Institute of Mathematical Statistics Monographs, Cambridge University Press, CambridgezbMATHGoogle Scholar
  6. Azzalini A, Capitanio A (2003) Distributions generated by perturbation of symmetry with emphasis on the multivariate skew t-distribution. J R Stat Soc B 61:367–389MathSciNetCrossRefzbMATHGoogle Scholar
  7. Azzalini A, Dalla-Vale A (1996) The multivariate skew-normal distribution. Biometrika 83:715–726MathSciNetCrossRefzbMATHGoogle Scholar
  8. Barndorff-Nielsen O (1978) Hyperbolic distributions and distributions on hyperbolae. Scand J Stat 5:151–157MathSciNetzbMATHGoogle Scholar
  9. Barndorff-Nielsen O, Blaesild P (1980) Hyperbolic distributions. In: Kotz S, Johnson NL, Read C (eds) Encyclopedia of statistical sciences, vol 3. Wiley, New YorkGoogle Scholar
  10. Barndorff-Nielsen O, Halgreen C (1977) Infinite divisibility of the hyperbolic and generalized inverse Gaussian distributions. Zeitschrift für Wahrscheinlichkeitstheorie und verwandte Gebiete 38:309–311MathSciNetCrossRefzbMATHGoogle Scholar
  11. Branco MD, Dey DK (2001) A general class of multivariate skew-elliptical distributions. J Multivar Anal 79:99–113MathSciNetCrossRefzbMATHGoogle Scholar
  12. Brooks SP (2002) Discussion on the paper by Spiegelhalter, Best, Carlin, and van der Linde. J R Stat Soc Ser B Stat Methodol 64(4):616–618Google Scholar
  13. Carlin BP, Louis TA (2001) Bayes and empirical Bayes methods for data analysis, 2nd edn. Chapman & Hall/CRC, Boca RatonzbMATHGoogle Scholar
  14. Chhikara RS, Folks JL (1989) The inverse Gaussian distribution. Marcel Dekker, New YorkzbMATHGoogle Scholar
  15. Dunson D (2010) Random effect and latent variable model selection. Lecture notes in statistics. Springer, New YorkzbMATHGoogle Scholar
  16. Forbes F, Wraith D (2014) Clustering using skewed multivariate heavy tailed distributions with flexible tail behaviour. Stat Comput 24(6):971–984MathSciNetCrossRefzbMATHGoogle Scholar
  17. Fruwirth-Schnatter S, Tuchler R, Otter T (2004) Bayesian analysis of the heterogeneity model. J Bus Econ Stat 22:2–15MathSciNetCrossRefGoogle Scholar
  18. Gamerman D (1997) Markov chain Monte Carlo: stochastic simulation for Bayesian inference. Chapman & Hill, LondonzbMATHGoogle Scholar
  19. Gelfand AE, Smith AFM (1990) Sampling based approaches to calculating marginal densities. J Am Stat Assoc 85:398–409MathSciNetCrossRefzbMATHGoogle Scholar
  20. Gelfand AE, Dey DK, Chang H (1992) Model determination using predictive distributions with implementation via sampling-based methods. In: Bayesian statistics, vol. 4. Peñíscola, Oxford Univ. Press, New York, pp 147–167Google Scholar
  21. Gelman A, Rubin DB (1992) Inference from iterative simulation using multiple sequences (with discussion). Stat Sci 7:457–511CrossRefzbMATHGoogle Scholar
  22. Good IJ (1953) The population frequencies of species and the estimation of population parameters. Biometrika 40:237–260MathSciNetCrossRefzbMATHGoogle Scholar
  23. Ho HJ, Lin TI (2010) Robust linear mixed models using the skew t distribution with application to schizophrenia data. Biom J 52(4):449–469MathSciNetCrossRefzbMATHGoogle Scholar
  24. Hogan JW, Laird NM (1997) Mixture models for the joint distribution of repeated measures and event times. Stat Med 16:239–258CrossRefGoogle Scholar
  25. Jara A, Quintana F, Martin ES (2008) Linear mixed models with skew-elliptical distributions: a Bayesian approach. Comput Stat Data Anal 52:5033–5045MathSciNetCrossRefzbMATHGoogle Scholar
  26. Johnson NL, Kotz S, Balakrishnan N (1994) Continuous univariate distributions, vol 1. Wiley, New YorkzbMATHGoogle Scholar
  27. Jones MC (2015) On families of distributions with shape parameters. Int Stat Rev 83(2):175–192MathSciNetCrossRefGoogle Scholar
  28. Jørgensen B (1982) Statistical properties of the generalized inverse Gaussian distribution. Springer, New YorkCrossRefzbMATHGoogle Scholar
  29. Lachos VH, Bolfarine H, Montenegro LC (2007) Influence diagnostics for skew-normal linear mixed models. Sankhya Indian Stat Inst 69:648–670MathSciNetzbMATHGoogle Scholar
  30. Lachos VH, Dey KD, Cancho VG (2009) Robust linear mixed models with skew-normal independent distributions from a bayesian perspective. J Stat Plan Inference 139:4098–4110MathSciNetCrossRefzbMATHGoogle Scholar
  31. Lachos VH, Ghosh P, Arellano-Valle RB (2010) Likelihood based inference for skew-normal independent linear mixed models. Statistica Sinica 20:303–322MathSciNetzbMATHGoogle Scholar
  32. Laird NM, Ware JH (1982) Random effects models for longitudinal data. Biometrics 38:963–974CrossRefzbMATHGoogle Scholar
  33. Lange KL, Sinsheimer JS (1993) Normal/independent distributions and their application in robust regression. J Comput Gr Stat 2:175–198MathSciNetGoogle Scholar
  34. Lange KL, Little RJA, Taylor JMG (1989) Robust statistical modeling using t-distributions. J Am Stat Assoc 84:881–896MathSciNetGoogle Scholar
  35. Lapierre YD, Nai NV, Chauinard G, Awad AG, Saxena B, James B, McClure DJ, Bakish D, Max P, Manchanda R, Beaudry P, Bloom D, Rotstein E, Ancill R, Sandor P, Sladen-Dew N, Durand C, Chandrasena R, Horn E, Elliot D, Das M, Ravindra A, Matsos G (1990) A controlled dose ranging study of remoxipride and haloperidol in schizophrenia: a Canadian multicentre trial. Acta Psychiatric Scandinavica 82:72–76CrossRefGoogle Scholar
  36. Lee SX, McLachlan GJ (2016) Finite mixtures of canonical fundamental skew t-distributions: the unification of the restricted and unrestricted skew-t mixture models. Stat Comput 26:573–589MathSciNetCrossRefzbMATHGoogle Scholar
  37. Lee JC, Lin TI, Lee KJ, Hsu YL (2005) Bayesian analysis of Box–Cox transformed linear mixed models with ARMA(p, q) dependence. J Stat Plan Inference 133:435–451MathSciNetCrossRefzbMATHGoogle Scholar
  38. Lin TI, Lee JC (2003) On modeling data from degradation sample paths over time. Aust N Z J Stat 45:257–270MathSciNetCrossRefzbMATHGoogle Scholar
  39. Lin TI, Lee JC (2006) A robust approach to t linear mixed models applied to multiple sclerosis data. Stat Med 25:1397–1412MathSciNetCrossRefGoogle Scholar
  40. Lin TI, Lee JC (2007) Bayesian analysis of hierarchical linear mixed modeling using the multivariate t distribution. J Stat Plan Inference 137:484–495MathSciNetCrossRefzbMATHGoogle Scholar
  41. Liu FG, Han B, Zhao X, Lin Q (2016) A comparison of frequentist and Bayesian model approaches for missing data analysis: case study with a schizophrenia clinical trial. Stat Biopharm Res 8(1):116–127CrossRefGoogle Scholar
  42. Maleki M, Mahmoudi MR (2017) Two-piece location-scale distributions based on scale mixtures of normal family. Commun Stat Theory Methods 46(24):12356–12369MathSciNetCrossRefzbMATHGoogle Scholar
  43. Modholkar GS, Hutson AD (2000) The epsilon-skew-normal distribution for analyzing near-normal data. J Stat Plan Inference 83:291–309MathSciNetCrossRefzbMATHGoogle Scholar
  44. Mueller P, Quintana FA, Rosner GL, Maitland ML (2014) Bayesian inference for longitudinal data with non-parametric treatment effects. Biostatistics 15(2):341–352CrossRefGoogle Scholar
  45. NIMBLE Development Team (2017) NIMBLE: an R package for programming with BUGS models, Version 0.6-10. http://r-nimble.org
  46. Pinheiro J, Liu C, Wu YN (2001) Efficient algorithms for robust estimation in linear mixed-effects models using the multivariate-t distribution. J Comput Gr Stat 10:249–276MathSciNetCrossRefGoogle Scholar
  47. Plummer M (2017) JAGS version 4.3.0 user manual. 28 June 2017. http://mcmc-jags.sourceforge.net/
  48. R Core Team (2017) R: a language and environment for statistical computing. R Foundation for Statistical Computing, Vienna, Austria. URL https://www.R-project.org/
  49. Robert CP (2015) The expected demise of the Bayes factor. J Math Psychol.  https://doi.org/10.1016/j.jmp.2015.08.002 zbMATHGoogle Scholar
  50. Robert CP, Wraith D (2009) Computational methods for Bayesian model choice. In: Bayesian inference and maximum entropy methods in science and engineering: The 29th international workshop on Bayesian inference and maximum entropy methods in science and engineering. AIP Conference Proceedings, 1193, pp 251–262Google Scholar
  51. Sahu SK, Dey DK, Branco MD (2003) A new class of multivariate skew distributions with applications to Bayesian regression models. Can J Stat 31(2):129–150MathSciNetCrossRefzbMATHGoogle Scholar
  52. Seshadri V (1993) The inverse Gaussian distribution: a case study in exponential families. Oxford University Press, New YorkGoogle Scholar
  53. Shih WJ, Quan H (1997) Testing for treatment difference with dropouts present in clinical trials—a composite approach. Stat Med 16:1225–1239CrossRefGoogle Scholar
  54. Stan Development Team (2017) The Stan Core Library, version 2.17.0. http://mc-stan.org
  55. Verbeke G, Lesaffre E (1996) A linear mixed-effects model with heterogeneity in the random-effects population. J Am Stat Assoc 91:217–221CrossRefzbMATHGoogle Scholar
  56. Verbeke G, Molenberghs G (2000) Linear mixed models for longitudinal data. Springer, New YorkzbMATHGoogle Scholar
  57. Vilca F, Balakrishnan N, Zeller CB (2014) Multivariate skew-normal generalized hyperbolic distribution and its properties. J Multivar Anal 128:73–85MathSciNetCrossRefzbMATHGoogle Scholar
  58. Wraith D, Forbes F (2015) Location and scale mixtures of Gaussians with flexible tail behaviour: properties, inference and application to multivariate clustering. Comput Stat Data Anal 90:61–73MathSciNetCrossRefzbMATHGoogle Scholar
  59. Wu H, Zhang J (2006) Nonparametric regression methods for longitudinal data analysis. Wiley, New-YorkzbMATHGoogle Scholar

Copyright information

© Sociedad de Estadística e Investigación Operativa 2018

Authors and Affiliations

  1. 1.Department of StatisticsShiraz UniversityShirazIran
  2. 2.Institute of Health and Biomedical Innovation (IHBI)Queensland University of Technology (QUT)BrisbaneAustralia
  3. 3.Department of StatisticsUniversidad Católica de ChileSantiagoChile

Personalised recommendations