SAS/STAT Software

BGLIMM Procedure

The BGLIMM procedure is a high-performance, sampling-based procedure that provides Bayesian inference for generalized linear mixed models (GLMMs). PROC BGLIMM uses syntax similar to that of PROC MIXED and PROC GLIMMIX in specifying a GLMM. The following are highlights of the BGLIMM procedure's features:

  • GLMMs with univariate or multivariate random effects
  • nested or non-nested hierarchical models
  • repeated measurements models (balanced or unbalanced data) with normal data
  • suite of covariance structures for random effects and residuals
  • built-in prior distributions for regression coefficients and covariance parameters
  • model heterogeneity in covariance structures
  • produce estimate and credible intervals for estimate linear combination of effects
  • support for missing completely at random (MCAR) and missing at random (MAR) approaches in modeling missing data
  • works with the postprocessing autocall macros that are designed for Bayesian posterior samples
  • provides a variety of Markov chain convergence diagnostics
  • creates an output data set that contains the posterior samples of all parameters
  • supports a CLASS statement for specifying classification variables
  • supports BY group processing
  • automatically produces graphs by using ODS Graphics
  • multithreaded

For further details see the SAS/STAT User's Guide: The BGLIMM Procedure ( PDF | HTML )