The HPSVM Procedure

PROC HPSVM Statement

PROC HPSVM DATA=SAS-data-set <options>;

The PROC HPSVM statement invokes the procedure. You must specify the following argument:

DATA=SAS-data-set: specifies the input SAS data set or database table to be used by PROC HPSVM.

You can also specify the following options:

C=number

specifies the penalty value. The value of number must be a real number greater than 0.

By default, C=1.0.

MAXITER=number

specifies the maximum number of iterations before the process stops. You must specify a positive integer for number.

This option applies only to the interior point method. By default, MAXITER=25.

METHOD=IPOINT | ACTIVESET

specifies which method to use during training. You can specify following values:

ACTIVESET: uses the active-set method during training.
IPOINT: uses the interior point method during training.

By default, METHOD=IPOINT.

NOMISS

requests that missing values for categorical variables be used. In this case, the missing value is treated as a special level of the categorical variables.

By default, all missing value observations are dropped during the training process.

NOPRINT

suppresses the generation of ODS outputs. If you specify this option, no ODS tables are generated.

NOSCALE

suppresses scaling during model fitting. (By default, PROC HPSVM scales the input variables to be between 0 and 1 inclusive.)

TOLERANCE=number

specifies the minimal tolerance at which the iteration stops. The tolerance number must equal to or greater than 1.0E–12.

By default, TOLERANCE=0.000001.

Last updated: May 25, 2022