Communications for Statistical Applications and Methods

The Korean Statistical Society (한국통계학회)
권호 표지
DOI QR코드

DOI QR Code

Bayesian Modeling of Random Effects Covariance Matrix for Generalized Linear Mixed Models

  • Lee, Keunbaik (Department of Statistics, Sungkyunkwan University)
  • Received : 2013.04.14
  • Accepted : 2013.05.21
  • Published : 2013.05.31
Download PDF
⟨ Previous Next ⟩

Abstract

Generalized linear mixed models(GLMMs) are frequently used for the analysis of longitudinal categorical data when the subject-specific effects is of interest. In GLMMs, the structure of the random effects covariance matrix is important for the estimation of fixed effects and to explain subject and time variations. The estimation of the matrix is not simple because of the high dimension and the positive definiteness; subsequently, we practically use the simple structure of the covariance matrix such as AR(1). However, this strong assumption can result in biased estimates of the fixed effects. In this paper, we introduce Bayesian modeling approaches for the random effects covariance matrix using a modified Cholesky decomposition. The modified Cholesky decomposition approach has been used to explain a heterogenous random effects covariance matrix and the subsequent estimated covariance matrix will be positive definite. We analyze metabolic syndrome data from a Korean Genomic Epidemiology Study using these methods.

Keywords

References

  1. Breslow, N. E. and Clayton, D. G. (1993). Approximate inference in generalized linear mixed models, Journal of the American Statistical Association, 88, 125-134.
  2. Celeux, G., Forbes, F., Robert, C. P. and Titterington, D. M. (2006). Deviance information criteria for missing data models, Bayesian Analysis, 1, 651-674. https://doi.org/10.1214/06-BA122
  3. Daniels, M. J. and Hogan, J.W. (2008). Missing Data in Longitudinal Studies: Strategies for Bayesian Modeling and Sensitivity Analysis, Chapman & Hall/CRC.
  4. Daniels, J. M. and Pourahmadi, M. (2002). Bayesian analysis of covariance matrices and dynamic models for longitudinal data, Biometrika, 89, 553-566. https://doi.org/10.1093/biomet/89.3.553
  5. Daniels, J. M. and Zhao, Y. D. (2003). Modelling the random effects covariance matrix in longitudinal data, Statistics in Medicine, 22, 1631-1647. https://doi.org/10.1002/sim.1470
  6. Heagerty, P. J. and Kurland, B. F. (2001). Misspecified maximum likelihood estimates and generalized linear mixed models, Biometrika, 88, 973-985. https://doi.org/10.1093/biomet/88.4.973
  7. Kim, J., Kim, E., Yi, H., Joo, S., Shin, K., Kim, J., Kim, K. and Shin, C. (2006). Short-term incidence rate of hypertension in Korea middle-aged adults, Journal of Hypertension, 24, 2177-2182. https://doi.org/10.1097/01.hjh.0000249694.81241.7c
  8. Lee, K., Yoo, J. K., Lee, J. and Hagan, J. (2012). Modeling the random effects covariance matrix for the generalized linear mixed models, Computational Statistics & Data Analysis, 56, 1545-1551. https://doi.org/10.1016/j.csda.2011.09.011
  9. Pan, J. and Mackenzie, G. (2003). On modelling mean-covariance structures in longitudinal studies, Biometrika, 90, 239-244. https://doi.org/10.1093/biomet/90.1.239
  10. Pan, J. and Mackenzie, G. (2006). Regression models for covariance structures in longitudinal studies, Statistical Modelling, 6, 43-57. https://doi.org/10.1191/1471082X06st105oa
  11. Pourahmadi, M. (1999). Joint mean-covariance models with applications to longitudinal data: Unconstrained parameterisation, Biometrika, 86, 677-690. https://doi.org/10.1093/biomet/86.3.677
  12. Pourahmadi, M. (2000). Maximum likelihood estimation of generalized linear models for multivariate normal covariance matrix, Biometrika, 87, 425-435. https://doi.org/10.1093/biomet/87.2.425
  13. Pourahmadi, M. and Daniels, M. J. (2002). Dynamic conditionally linear mixed models for longitudinal data, Biometrics, 58, 225-231. https://doi.org/10.1111/j.0006-341X.2002.00225.x
  14. Spiegelhalter, D. J., Best, N. G., Carlin, B. P. and Van Der Linde, A. (2002). Bayesian measures of model complexity and fit, Journal of the Royal Statistical Society, Series B: Statistical Methodology, 64, 583-616. https://doi.org/10.1111/1467-9868.00353

Cited by

  1. A marginalized multilevel model for bivariate longitudinal binary data 2016, https://doi.org/10.1007/s00362-016-0840-1
  2. ARMA Cholesky factor models for the covariance matrix of linear models vol.115, 2017, https://doi.org/10.1016/j.csda.2017.05.001
  3. Bayesian modeling of random effects precision/covariance matrix in cumulative logit random effects models vol.24, pp.1, 2017, https://doi.org/10.5351/CSAM.2017.24.1.081
  4. Survey of Models for Random Effects Covariance Matrix in Generalized Linear Mixed Model vol.28, pp.2, 2015, https://doi.org/10.5351/KJAS.2015.28.2.211
  5. Dynamic linear mixed models with ARMA covariance matrix vol.23, pp.6, 2016, https://doi.org/10.5351/CSAM.2016.23.6.575
  6. Hurdle Model for Longitudinal Zero-Inflated Count Data Analysis vol.27, pp.6, 2014, https://doi.org/10.5351/KJAS.2014.27.6.923
  7. Bayesian Cholesky factor models in random effects covariance matrix for generalized linear mixed models vol.80, 2014, https://doi.org/10.1016/j.csda.2014.06.016
  8. Negative binomial loglinear mixed models with general random effects covariance matrix vol.25, pp.1, 2018, https://doi.org/10.29220/CSAM.2018.25.1.061