×

Fast and scalable computations for Gaussian hierarchical models with intrinsic conditional autoregressive spatial random effects. (English) Zbl 07422757

Summary: Fast algorithms are developed for Bayesian analysis of Gaussian hierarchical models with intrinsic conditional autoregressive (ICAR) spatial random effects. To achieve computational speed-ups, first a result is proved on the equivalence between the use of an improper CAR prior with centering on the fly and the use of a sum-zero constrained ICAR prior. This equivalence result then provides the key insight for the algorithms, which are based on rewriting the hierarchical model in the spectral domain. The two novel algorithms are the Spectral Gibbs Sampler (SGS) and the Spectral Posterior Maximizer (SPM). Both algorithms are based on one single matrix spectral decomposition computation. After this computation, the SGS and SPM algorithms scale linearly with the sample size. The SGS algorithm is preferable for smaller sample sizes, whereas the SPM algorithm is preferable for sample sizes large enough for asymptotic calculations to provide good approximations. Because the matrix spectral decomposition needs to be computed only once, the SPM algorithm has computational advantages over algorithms based on sparse matrix factorizations (which need to be computed for each value of the random effects variance parameter) in situations when many models need to be fitted. Three simulation studies are performed: the first simulation study shows improved performance in computational speed in estimation of the SGS algorithm compared to an algorithm that uses the spectral decomposition of the precision matrix; the second simulation study shows that for model selection computations with 10 regressors and sample sizes varying from 49 to 3600, when compared to the current fastest state-of-the-art algorithm implemented in the R package INLA, SPM computations are 550 to 1825 times faster; the third simulation study shows that, when compared to default INLA settings, SGS and SPM combined with reference priors provide much more adequate uncertainty quantification. Finally, the application of the novel SGS and SPM algorithms is illustrated with a spatial regression study of county-level median household income for 3108 counties in the contiguous United States in 2017.

MSC:

62-XX Statistics
PDFBibTeX XMLCite
Full Text: DOI

References:

[1] Banerjee, S.; Carlin, B. P.; Gelfand, A. E., Hierarchical Modeling and Analysis for Spatial Data (2014), CRC Press: CRC Press Boca Raton, FL
[2] Berger, J., The case for objective Bayesian analysis, Bayesian Anal., 1, 385-402 (2006) · Zbl 1331.62042
[3] Berger, J. O.; de Oliveira, V.; Sansó, B., Objective Bayesian analysis of spatially correlated data, J. Am. Stat. Assoc., 96, 1361-1374 (2001) · Zbl 1051.62095
[4] Bernardinelli, L.; Clayton, D.; Montomoli, C., Bayesian estimates of disease maps: how important are priors?, Stat. Med., 14, 2411-2431 (1995)
[5] Besag, J., Spatial interaction and the statistical analysis of lattice systems, J. R. Stat. Soc., Ser. B, Stat. Methodol., 192-236 (1974) · Zbl 0327.60067
[6] Besag, J.; Green, P.; Higdon, D.; Mengersen, K., Bayesian computation and stochastic systems (with discussion), Stat. Sci., 10, 3-66 (1995) · Zbl 0955.62552
[7] Besag, J.; Kooperberg, C., On conditional and intrinsic autoregressions, Biometrika, 82, 733-746 (1995) · Zbl 0899.62123
[8] Besag, J.; York, J.; Mollié, A., Bayesian image restoration, with two applications in spatial statistics, Ann. Inst. Stat. Math., 43, 1-20 (1991) · Zbl 0760.62029
[9] Best, N. G.; Arnold, R. A.; Thomas, A.; Waller, L. A.; Conlon, E. M., Bayesian models for spatially correlated disease and exposure data, (Bayesian Statistics 6: Proceedings of the Sixth Valencia International Meeting (1999), Oxford University Press), 131 · Zbl 0973.62020
[10] Bivand, R. S.; Pebesma, E. J.; Gómez-Rubio, V.; Pebesma, E. J., Applied Spatial Data Analysis with R (2013), Springer: Springer New York · Zbl 1269.62045
[11] Breslow, N. E.; Clayton, D. G., Approximate inference in generalized linear mixed models, J. Am. Stat. Assoc., 88, 9-25 (1993) · Zbl 0775.62195
[12] Celeux, G.; Forbes, F.; Robert, C. P.; Titterington, D. M., Deviance information criteria for missing data models, Bayesian Anal., 1, 651-673 (2006) · Zbl 1331.62329
[13] Clayton, D.; Kaldor, J., Empirical Bayes estimates of age-standardized relative risks for use in disease mapping, Biometrics, 671-681 (1987)
[14] Crainiceanu, C. M.; Ruppert, D., Likelihood ratio tests in linear mixed models with one variance component, J. R. Stat. Soc., Ser. B, Stat. Methodol., 66, 165-185 (2004) · Zbl 1061.62027
[15] De Oliveira, V.; Ferreira, M. A.R., Maximum likelihood and restricted maximum likelihood estimation for a class of Gaussian Markov random fields, Metrika, 74, 167-183 (2011) · Zbl 1230.62106
[16] Efron, B., Frequentist accuracy of Bayesian estimates, J. R. Stat. Soc., Ser. B, Stat. Methodol., 77, 617 (2015) · Zbl 1414.62089
[17] Ferreira, M. A.R., The limiting distribution of the Gibbs sampler for the intrinsic conditional autoregressive model, Braz. J. Probab. Stat., 33, 734-744 (2019) · Zbl 1432.62299
[18] Ferreira, M. A.R.; Bertolde, A. I.; Holan, S., Analysis of economic data with multi-scale spatio-temporal models, (O’Hagan, A.; West, M., Handbook of Applied Bayesian Analysis (2010), Oxford University Press: Oxford University Press Oxford), 295-318
[19] Ferreira, M. A.R.; De Oliveira, V., Bayesian reference analysis for Gaussian Markov random fields, J. Multivar. Anal., 98, 789-812 (2007) · Zbl 1118.62100
[20] Ferreira, M. A.R.; Holan, S. H.; Bertolde, A. I., Dynamic multiscale spatio-temporal models for Gaussian areal data, J. R. Stat. Soc., Ser. B, Stat. Methodol., 73, 663-688 (2011) · Zbl 1228.62116
[21] Ferreira, M. A.R.; Salazar, E., Bayesian reference analysis for exponential power regression models, J. Stat. Distributions Appl., 1, 1-12 (2014) · Zbl 1327.62142
[22] Ferreira, M. A.R.; Suchard, M. A., Bayesian analysis of elapsed times in continuous-time Markov chains, Can. J. Stat., 36, 355-368 (2008) · Zbl 1153.62019
[23] Fonseca, T. C.O.; Ferreira, M. A.R.; Migon, H. S., Objective Bayesian analysis for the Student-t regression model, Biometrika, 95, 325-333 (2008) · Zbl 1400.62260
[24] Fonseca, T. C.O.; Migon, H. S.; Ferreira, M. A.R., Bayesian analysis based on the Jeffreys prior for the hyperbolic distribution, Braz. J. Probab. Stat., 26, 327-343 (2012) · Zbl 1319.62060
[25] Freni-Sterrantino, A.; Ventrucci, M.; Rue, H., A note on intrinsic conditional autoregressive models for disconnected graphs, Spat. Spatio-Tempor. Epidemiol., 26, 25-34 (2018)
[26] Gelman, A.; Rubin, D. B., Inference from iterative simulation using multiple sequences, Stat. Sci., 7, 457-511 (1992) · Zbl 1386.65060
[27] He, D.; Sun, D.; He, L., Objective bayesian analysis for the Student-t linear regression, Bayesian Anal., 16, 129-145 (2021) · Zbl 1475.62206
[28] Hodges, J. S.; Carlin, B. P.; Fan, Q., On the precision of the conditionally autoregressive prior in spatial models, Biometrics, 59, 317-322 (2003) · Zbl 1210.62128
[29] Kang, H. M.; Zaitlen, N. A.; Wade, C. M.; Kirby, A.; Heckerman, D.; Daly, M. J.; Eskin, E., Efficient control of population structure in model organism association mapping, Genetics, 178, 1709-1723 (2008)
[30] Keefe, M. J.; Ferreira, M. A.R.; Franck, C. T., On the formal specification of sum-zero constrained intrinsic conditional autoregressive models, Spat. Stat., 24, 54-65 (2018)
[31] Keefe, M. J.; Ferreira, M. A.R.; Franck, C. T., Objective Bayesian analysis for Gaussian hierarchical models with intrinsic conditional autoregressive priors, Bayesian Anal., 14, 181-209 (2019) · Zbl 1409.62187
[32] Lavine, M. L.; Hodges, J. S., On rigorous specification of ICAR models, Am. Stat., 66, 42-49 (2012) · Zbl 07648997
[33] Lee, D., CARBayes: an R package for Bayesian spatial modeling with conditional autoregressive priors, J. Stat. Softw., 55, 1-24 (2013)
[34] Magnus, J. R.; Neudecker, H., Matrix Differential Calculus with Applications in Statistics and Econometrics (1999), Wiley: Wiley Chichester · Zbl 0912.15003
[35] Martins, T. G.; Simpson, D.; Lindgren, F.; Rue, H., Bayesian computing with INLA: new features, Comput. Stat. Data Anal., 67, 68-83 (2013) · Zbl 1471.62135
[36] Plummer, M.; Best, N.; Cowles, K.; Vines, K., Coda: convergence diagnosis and output analysis for mcmc, R News, 6, 7-11 (2006)
[37] Porter, E.M., Franck, C.T., Ferreira, M.A.R., 2021. Objective Bayesian model selection for spatial hierarchical models with intrinsic conditional autoregressive priors. Submitted for publication.
[38] Porter, E.M., Keefe, M.J., Franck, C.T., Ferreira, M.A.R., 2019. Ref.ICAR: Objective Bayes Intrinsic Conditional Autoregressive Model for Areal Data. R package version 1.0.
[39] R: A Language and Environment for Statistical Computing (2014), R Foundation for Statistical Computing: R Foundation for Statistical Computing Vienna, Austria
[40] Rue, H., 2021. Personal communication.
[41] Rue, H.; Held, L., Gaussian Markov Random Fields: Theory and Applications (2005), CRC Press · Zbl 1093.60003
[42] Rue, H.; Martino, S.; Chopin, N., Approximate Bayesian inference for latent Gaussian models by using integrated nested Laplace approximations (with discussion), J. R. Stat. Soc., Ser. B, Stat. Methodol., 71, 319-392 (2009) · Zbl 1248.62156
[43] Salazar, E.; Ferreira, M. A.R.; Migon, H. S., Objective bayesian analysis for exponential power regression models, Sankhya, Ser. B, 74, 107-125 (2012) · Zbl 1257.62029
[44] Sørbye, S. H.; Rue, H., Scaling intrinsic Gaussian Markov random field priors in spatial modelling, Spat. Stat., 8, 39-51 (2014)
[45] Spiegelhalter, D. J.; Best, N. G.; Carlin, B. P.; Van Der Linde, A., Bayesian measures of model complexity and fit, J. R. Stat. Soc., Ser. B, Stat. Methodol., 64, 583-639 (2002) · Zbl 1067.62010
[46] Sun, D.; Berger, J. O., Objective bayesian analysis for the multivariate normal model, (Bernardo, J. M.; Bayarri, M.; Berger, J.; Dawid, A.; Heckerman, D.; Smith, A.; West, M., Bayesian Statistics 8 (2007), Oxford University Press: Oxford University Press Oxford), 525-547 · Zbl 1252.62030
[47] Thomas, A.; Best, N.; Lunn, D.; Arnold, R.; Spiegelhalter, D., GeoBugs User Manual (2004), Medical Research Council Biostatistics Unit.: Medical Research Council Biostatistics Unit. Cambridge
[48] Watanabe, S., Asymptotic equivalence of Bayes cross validation and widely applicable information criterion in singular learning theory, J. Mach. Learn. Res., 11, 3571-3594 (2010) · Zbl 1242.62024
This reference list is based on information provided by the publisher or from digital mathematics libraries. Its items are heuristically matched to zbMATH identifiers and may contain data conversion errors. In some cases that data have been complemented/enhanced by data from zbMATH Open. This attempts to reflect the references listed in the original paper as accurately as possible without claiming completeness or a perfect matching.