The GLIMMIX Procedure
-
Overview
-
Getting Started
-
SyntaxPROC GLIMMIX StatementBY StatementCLASS StatementCODE StatementCONTRAST StatementCOVTEST StatementEFFECT StatementESTIMATE StatementFREQ StatementID StatementLSMEANS StatementLSMESTIMATE StatementMODEL StatementNLOPTIONS StatementOUTPUT StatementPARMS StatementRANDOM StatementSLICE StatementSTORE StatementWEIGHT StatementProgramming StatementsUser-Defined Link or Variance Function
-
DetailsGeneralized Linear Models TheoryGeneralized Linear Mixed Models TheoryGLM Mode or GLMM ModeStatistical Inference for Covariance ParametersDegrees of Freedom MethodsEmpirical Covariance ("Sandwich") EstimatorsExploring and Comparing Covariance MatricesProcessing by SubjectsRadial Smoothing Based on Mixed ModelsOdds and Odds Ratio EstimationParameterization of Generalized Linear Mixed ModelsResponse-Level Ordering and ReferencingComparing the GLIMMIX and MIXED ProceduresSingly or Doubly Iterative FittingDefault Estimation TechniquesDefault OutputNotes on Output StatisticsODS Table NamesODS Graphics
-
ExamplesBinomial Counts in Randomized BlocksMating Experiment with Crossed Random EffectsSmoothing Disease Rates; Standardized Mortality RatiosQuasi-likelihood Estimation for Proportions with Unknown DistributionJoint Modeling of Binary and Count DataRadial Smoothing of Repeated Measures DataIsotonic Contrasts for Ordered AlternativesAdjusted Covariance Matrices of Fixed EffectsTesting Equality of Covariance and Correlation MatricesMultiple Trends Correspond to Multiple Extrema in Profile LikelihoodsMaximum Likelihood in Proportional Odds Model with Random EffectsFitting a Marginal (GEE-Type) ModelResponse Surface Comparisons with Multiplicity AdjustmentsGeneralized Poisson Mixed Model for Overdispersed Count DataComparing Multiple B-SplinesDiallel Experiment with Multimember Random EffectsLinear Inference Based on Summary DataWeighted Multilevel Model for Survey Data
- References
Most models fit with the GLIMMIX procedure typically have one or more nonlinear parameters. Estimation requires nonlinear optimization methods. You can control the optimization through options in the NLOPTIONS statement.
Several estimation methods of the GLIMMIX procedure (METHOD= RSPL, MSPL, RMPL, MMPL) are doubly iterative in the following sense. The generalized linear mixed model is approximated by a linear mixed model based on current values of the covariance parameter estimates. The resulting linear mixed model is then fit, which is itself an iterative process (with some exceptions). On convergence, new covariance parameters and fixed-effects estimates are obtained and the approximated linear mixed model is updated. Its parameters are again estimated iteratively. It is thus reasonable to refer to outer and inner iterations. The outer iterations involve the repeated updates of the linear mixed models, and the inner iterations are the iterative steps that lead to parameter estimates in any given linear mixed model. The NLOPTIONS statement controls the inner iterations. The outer iteration behavior can be controlled with options in the PROC GLIMMIX statement, such as the MAXLMMUPDATE= , PCONV= , and ABSPCONV= options. If the estimation method involves a singly iterative approach, then there is no need for the outer cycling and the model is fit in a single optimization controlled by the NLOPTIONS statement (see the section Singly or Doubly Iterative Fitting).
The syntax and options of the NLOPTIONS statement are described in the section NLOPTIONS Statement in Chapter 19: Shared Concepts and Topics.
Note that in a GLMM with pseudo-likelihood estimation, specifying TECHNIQUE=NONE has the same effect as specifying the NOITER option in the PARMS statement. If you estimate the parameters by METHOD= LAPLACE or METHOD= QUAD , TECHNIQUE=NONE applies to the optimization after starting values have been determined.
The GLIMMIX procedure applies the default optimization technique shown in Table 44.14, depending on your model.
Table 44.14: Default Techniques
|
Model Family |
Setting |
TECHNIQUE= |
|---|---|---|
|
GLM |
DIST=NORMAL |
NONE |
|
LINK=IDENTITY |
||
|
GLM |
otherwise |
NEWRAP |
|
GLMM |
PARMS NOITER, PL |
NONE |
|
GLMM |
binary data, PL |
NRRIDG |
|
GLMM |
otherwise |
QUANEW |