PROC GLIMMIX Additional Statements

Enter additional SAS PROC GLIMMIX options in this field to override default parameters.

Additional statements that you might want to specify include:

WHERE for subsetting rows of the input data set,
NLOPTIONS for convergence options, and
PARMS for performing a grid search of the likelihood or for specifying fixed values of the covariance parameters.

Refer to the PROC GLIMMIX documentation for appropriate syntax.

To specify one or more additional PROC GLIMMIX statements, use the following syntax:

Option

where:

Option is the PROC GLIMMIX option, and
a space is used to delimit individual conditions.

Examples of commonly used PROC GLIMMIX options are listed in the following table.

Option

Definition

WHERE

Used for subsetting rows in the input data set.

Refer to the SAS WHERE Statement documentation for additional details about WHERE statements.

NLOPTIONS

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 using the NLOPTIONS statement.

See the NLOPTIONS Statement documentation of more information

PARMS

The PARMS statement specifies initial values for the covariance or scale parameters, or it requests a grid search over several values of these parameters in generalized linear mixed models.

See the PARMS Statement documentation for more information.

Invalid Statements

By default, CLASS, MODEL, RANDOM, and OUTPUT statements are automatically generated to fit the radial basis machine, so do not specify any of these statements. Also, do not specify a BY statement, since the SAS code is currently not set up to handle BY groups.

To Specify One or More Additional PROC GLIMMIX Options:

8 Type specific PROC GLIMMIX options in the PROC GLIMMIX Statement Options field.

For example, to restrict the analysis to males only, type WHERE SEX=’M’in the text field, as shown below:

For Additional Information

Refer to the SAS PROC GLIMMIX documentation for more information.