The LMIXED Procedure
PROC LMIXED Features
The LMIXED procedure provides easy accessibility to numerous linear mixed models that are useful in many common statistical analyses.
Here are the main features of the LMIXED procedure:
The RANDOM statement supports many covariance structures, including variance components, compound symmetry, unstructured, AR(1), Toeplitz, factor analytic, and so on.
Both the MODEL statement and the RANDOM statement are supported for model specification, as in the MIXED procedure.
Inference features include standard errors and t tests for fixed and random effects.
A subject effect for blocking is supported.
Both REML and ML estimation methods are supported; they are implemented with a variety of optimization algorithms.
It handles unbalanced data.
Specialized dense and sparse matrix algorithms are provided.
The OUTPUT statement produces output data tables that contain predicted values, residuals, studentized residuals, confidence limits, and influence statistics.
The PARMS statement enables you to fit a linear mixed model that has known covariance values, or to set boundary values for the parameters.
Because the LMIXED procedure runs on SAS Cloud Analytic Services (CAS, see the section Using CAS Sessions and CAS Engine Librefs), it also does the following:
enables you to run on a cluster of machines that distribute the data and the computations
enables you to run in single-machine mode on CAS
exploits all the available cores and concurrent threads. For information about how the LMIXED procedure uses threads, see the section Multithreading in Chapter 3: Shared Concepts.