The MCMC Procedure
-
Overview
-
Getting Started
-
Syntax
-
DetailsHow PROC MCMC WorksBlocking of ParametersSampling MethodsTuning the Proposal DistributionDirect SamplingConjugate SamplingInitial Values of the Markov ChainsAssignments of ParametersStandard DistributionsUsage of Multivariate DistributionsSpecifying a New DistributionUsing Density Functions in the Programming StatementsTruncation and CensoringSome Useful SAS FunctionsMatrix Functions in PROC MCMCCreate Design MatrixModeling Joint LikelihoodRegenerating Diagnostics PlotsCaterpillar PlotAutocall Macros for PostprocessingGamma and Inverse-Gamma DistributionsPosterior Predictive DistributionHandling of Missing DataFunctions of Random-Effects ParametersFloating Point Errors and OverflowsHandling Error MessagesComputational ResourcesDisplayed OutputODS Table NamesODS Graphics
-
ExamplesSimulating Samples From a Known DensityBox-Cox TransformationLogistic Regression Model with a Diffuse PriorLogistic Regression Model with Jeffreys’ PriorPoisson RegressionNonlinear Poisson Regression ModelsLogistic Regression Random-Effects ModelNonlinear Poisson Regression Multilevel Random-Effects ModelMultivariate Normal Random-Effects ModelMissing at Random AnalysisNonignorably Missing Data (MNAR) AnalysisChange Point ModelsExponential and Weibull Survival AnalysisTime Independent Cox ModelTime Dependent Cox ModelPiecewise Exponential Frailty ModelNormal Regression with Interval CensoringConstrained AnalysisImplement a New Sampling AlgorithmUsing a Transformation to Improve MixingGelman-Rubin Diagnostics
- References
The BEGINCNST and ENDCNST statements define a block within which PROC MCMC processes the programming statements only during the setup stage of the simulation. You can use the BEGINCNST and ENDCNST statements to define constants or import data set variables into arrays. Storing data in arrays enables you to work with data that are not identically distributed (see the section Modeling Joint Likelihood) or to implement your own Markov chain sampler (see the section UDS Statement). You can also use the BEGINCNST and ENDCNST statements to assign initial values to the parameters (see the section Assignments of Parameters).
Whenever you have programming statements that calculate constants that do not need to be evaluated multiple times throughout the simulation, you should put them within the BEGINCNST and ENDCNST statements. Using these statements can reduce redundant processing. For example, you can assign a constant to a symbol or fill in an array with numbers:
array cnst[17];
begincnst;
offset = 17;
do i = 1 to 17;
cnst[i] = i * i;
end;
endcnst;
During the setup process, PROC MCMC evaluates the programming statements within the BEGINCNST/ENDCNST once for each observation in the data set and ignores the statements in the rest of the simulation.
Sometimes you might need to store variables, either from the current input data set or from a different data set, in arrays and use these arrays to specify your model. The READ_ARRAY function is convenient for that purpose.
The following two forms of the READ_ARRAY function are available:
where
-
rc returns 0 if the function is able to successfully read the data set.
-
data-set specifies the name of the data set from which the array data is read. The value specified for data-set must be a character literal or a variable that contains the member name (
libname.memname) of the data set to be read from. -
array specifies the PROC MCMC array variable into which the data is read. The value specified for array must be a local temporary array variable because the function might need to grow or shrink its size to accommodate the size of the data set.
-
col-name specifies optional names for the specific columns of the data set that are read. If specified, col-name must be a literal string enclosed in quotation marks. In addition, col-name cannot be a PROC MCMC variable. If column names are not specified, PROC MCMC reads all of the columns in the data set.
When SAS translates between an array and a data set, the array is indexed as [row,column].
The READ_ARRAY function attempts to dynamically resize the array to match the dimensions of the input data set. Therefore, the array must be dynamic; that is, the array must be declared with the /NOSYMBOLS option.
For examples that use the READ_ARRAY function, see Modeling Joint Likelihood, Time Independent Cox Model, and Implement a New Sampling Algorithm.