OpenLB
Open Library of Bioscience
Advanced Search
Statistics in medicine
06 30, 2022;41 (14): 2665-2687.

Bayesian semi-parametric modeling of covariance matrices for multivariate longitudinal data.

Georgios Papageorgiou
Author Information
  1. Georgios Papageorgiou: Department of Economics, Mathematics and Statistics, Birkbeck, University of London, London, UK. ORCID
PMID: 35699116 DOI: 10.1002/sim.9376

Abstract

The article develops marginal models for multivariate longitudinal responses. Overall, the model consists of five regression submodels, one for the mean and four for the covariance matrix, with the latter resulting by considering various matrix decompositions. The decompositions that we employ are intuitive, easy to understand, and they do not rely on any assumptions such as the presence of an ordering among the multivariate responses. The regression submodels are semi-parametric, with unknown functions represented by basis function expansions. We use spike-slap priors for the regression coefficients to achieve variable selection and function regularization, and to obtain parameter estimates that account for model uncertainty. An efficient Markov chain Monte Carlo algorithm for posterior sampling is developed. The simulation study presented investigates the gains that one may have when considering multivariate longitudinal analyses instead of univariate ones, and whether these gains can counteract the negative effects of missing data. We apply the methods on a highly unbalanced longitudinal dataset with four responses observed over a period of 20 years.

Keywords

Cholesky decomposition clustering model averaging semi-parametric regression variable selection

References

  1. Hamilton JD. Time Series Analysis. Princeton, NJ: Princeton University Press; 1994.
  2. Xu J, Mackenzie G. Modelling covariance structure in bivariate marginal models for longitudinal data. Biometrika. 2012;99:649-662.
  3. Kim C, Zimmerman DL. Unconstrained models for the covariance structure of multivariate longitudinal data. J Multivar Anal. 2012;107:104-118.
  4. Kohli P, Garcia TP, Pourahmadi M. Modeling the Cholesky factors of covariance matrices of multivariate longitudinal data. J Multivar Anal. 2016;145:87-100.
  5. Lee K, Cho H, Kwak M-S, Jang EJ. Estimation of covariance matrix of multivariate longitudinal data using modified Choleksky and hypersphere decompositions. Biometrics. 2020;76:75-86.
  6. Pourahmadi M. Joint mean-covariance models with applications to longitudinal data: unconstrained parameterisation. Biometrika. 1999;86:677-690.
  7. Daniels MJ, Pourahmadi M. Bayesian analysis of covariance matrices and dynamic models for longitudinal data. Biometrika. 2002;89:553-566.
  8. Zhang W, Leng C. A moving average Cholesky factor model in covariance modelling for longitudinal data. Biometrika. 2012;99:141-150.
  9. Feng S, Lian H, Xue L. A new nested Cholesky decomposition and estimation for the covariance matrix of bivariate longitudinal data. Comput Stat Data Anal. 2016;102:98-109.
  10. Chiu TYM, Leonard T, Tsui K-W. The matrix-logarithmic covariance model. J Am Stat Assoc. 1996;91:198-210.
  11. Pourahmadi M. Maximum likelihood estimation of generalised linear models for multivariate normal covariance matrix. Biometrika. 2000;87:425-435.
  12. Anderson TW. Asymptotically efficient estimation of covariance matrices with linear structure. Ann Stat. 1973;1:135-141.
  13. Barnard J, McCulloch R, Meng X-L. Modeling covariance matrices in terms of standard deviations and correlations, with application to shrinkage. Stat Sin. 2000;10:1281-1311.
  14. Pourahmadi M. Cholesky decompositions and estimation of a covariance matrix: orthogonality of variance-correlation parameters. Biometrika. 2007;94:1006-1013.
  15. Zhang W, Leng C, Tang CY. A joint modelling approach for longitudinal studies. J Royal Stat Soc Ser B (Stat Methodol). 2015;77:219-238.
  16. Daniels MJ, Kass RE. Nonconjugate Bayesian estimation of covariance matrices and its use in hierarchical models. J Am Stat Assoc. 1999;94:1254-1263.
  17. Liechty JC, Liechty MW, Müller P. Bayesian correlation estimation. Biometrika. 2004;91:1-14.
  18. George EI, McCulloch RE. Approaches for Bayesian variable selection. Stat Sin. 1997;7:339-373.
  19. Chan D, Kohn R, Nott D, Kirby C. Locally adaptive semiparametric estimation of the mean and variance functions in regression models. J Comput Graph Stat. 2006;15:915-936.
  20. Papageorgiou G. BNSP: an R package for fitting Bayesian semiparametric regression models and variable selection. R I Dent J. 2018;10:526-548.
  21. Papageorgiou G, Marshall BC. Bayesian semiparametric analysis of multivariate continuous responses, with variable selection. J Comput Graph Stat. 2020;29:896-909.
  22. Zellner A. On assessing prior distributions and Bayesian regression analysis with g-prior distributions. In: Goel P, Zellner A, eds. Bayesian Inference and Decision Techniques: Essays in Honor of Bruno de Finetti. New York: Elsevier Science Publishers; 1986:233-243.
  23. Roberts GO, Rosenthal JS. Optimal scaling for various Metropolis-Hastings algorithms. Stat Sci. 2001;16:351-367.
  24. Papageorgiou G. BNSP: Bayesian non- and semi-parametric model fitting; 2020. R package version 2.1.5.
  25. Leng C, Zhang W, Pan J. Semiparametric mean-covariance regression analysis for longitudinal data. J Am Stat Assoc. 2010;105:181-193.
  26. Lin H, Pan J. Nonparametric estimation of mean and covariance structures for longitudinal data. Can J Stat. 2013;41:557-574.
  27. Ferguson TS. A Bayesian analysis of some nonparametric problems. Ann Stat. 1973;1:209-230.
  28. Sethuraman J. A constructive definition of Dirichlet priors. Stat Sin. 1994;4:639-650.
  29. Liang F, Paulo R, Molina G, Clyde MA, Berger JO. Mixtures of g priors for Bayesian variable selection. J Am Stat Assoc. 2008;103:410-423.
  30. Zhang X, Boscardin JW, Belin TR. Sampling correlation matrices in bayesian models with correlated latent variables. J Comput Graph Stat. 2006;15:880-896.
  31. Liu X, Daniels MJ. A new algorithm for simulating a correlation matrix based on parameter expansion and reparameterization. J Comput Graph Stat. 2006;15:897-914.
  32. Zellner A. An efficient method of estimating seemingly unrelated regressions and tests for aggregation bias. J Am Stat Assoc. 1962;57:348-368.
  33. Letenneur L, Commenges D, Dartigues JF, Barberger-Gateau P. Incidence of dementia and Alzheimer's disease in elderly community residents of South-Western France. Int J Epidemiol. 1994;23:1256-1261.
  34. Proust-Lima C, Philipps V, Diakite A, Liquet B. lcmm: extended mixed models using latent classes and latent processes; 2020. R package version: 1.9.2.
  35. Philipps V, Amieva H, Andrieu S, et al. Normalized mini-mental state examination for assessing cognitive change in population-based brain aging studies. Neuroepidemiology. 2014;43:15-25.
  36. Plummer M, Best N, Cowles K, Vines K. CODA: convergence diagnosis and output analysis for MCMC. R News. 2006;6:7-11.
  37. Pan J, MacKenzie G. Regression models for covariance structures in longitudinal studies. Stat Model. 2006;6:43-57.
  38. Papageorgiou G. Restricted maximum likelihood estimation of joint mean-covariance models. Can J Stat. 2012;40:225-242.
  39. Pan J, Mackenzie G. On modelling mean-covariance structures in longitudinal studies. Biometrika. 2003;90:239-244.
  40. Azzalini A, Dalla VA. The multivariate skew-normal distribution. Biometrika. 1996;83:715-726.

MeSH Term

Bayes Theorem
Computer Simulation
Humans
Markov Chains
Monte Carlo Method
Multivariate Analysis
Journal Article

Word Cloud

Created with Highcharts 10.0.0multivariatelongitudinalregressionresponsesmodelsemi-parametricsubmodelsonefourcovariancematrixconsideringdecompositionsfunctionvariableselectiongainsdataarticledevelopsmarginalmodelsOverallconsistsfivemeanlatterresultingvariousemployintuitiveeasyunderstandrelyassumptionspresenceorderingamongunknownfunctionsrepresentedbasisexpansionsusespike-slappriorscoefficientsachieveregularizationobtainparameterestimatesaccountuncertaintyefficientMarkovchainMonteCarloalgorithmposteriorsamplingdevelopedsimulationstudypresentedinvestigatesmayanalysesinsteadunivariateoneswhethercancounteractnegativeeffectsmissingapplymethodshighlyunbalanceddatasetobservedperiod20yearsBayesianmodelingmatricesCholeskydecompositionclusteringaveraging

Similar Articles

Cited By