DOCUMENTATION
Sign In

Get access to My SAS, trials, communities and more.

Machine Learning Procedures
 

REGISTERMODEL Procedure

ASSESSMENT Statement

  • ASSESSMENT <options>;

  • ASSESS <options>;

The ASSESSMENT statement enables you to assess the model and register assessment statistics in SAS Model Manager. To enable assessment, the FUNCTION= option value in the PROC REGISTERMODEL statement must be PREDICTION or CLASSIFICATION. You must also specify the TARGET statement and provide the training data table by specifying the DATA= option in the PROC REGISTERMODEL statement.

You can specify the following options:

FREQ=variable

identifies a numeric variable in the input data table and the holdout data table that contains the frequency of occurrence of each observation. PROC REGISTERMODEL treats each observation as if it appears f times, where f is the value of the frequency variable for the observation. If f is not an integer, it is truncated to an integer. If f is less than 1 or missing, the observation is not used in the analysis. When you omit the FREQ= option, each observation is assigned a frequency of 1. This variable is optional in the holdout data table. If the holdout data table does not contain this variable, each observation in the holdout data table is assigned a frequency of 1.

HOLDOUT=libref.data-table

names the holdout data table for assessment. libref.data-table is a two-level name, where libref refers to the library, and data-table specifies the name of the input data table. For more information about this two-level name, see the DATA= option and the section Using CAS Sessions and CAS Engine Librefs.

NCUTS=10 | 20 | 50 | 100 | 500 | 1000

specifies the number of cuts to use in the ROC calculation. This option is ignored when the level of the target variable is interval.

By default, NCUTS=100.

PARTITIONVAR=variable(<TEST='value'> <TRAIN='value'> <VALIDATE='value'> )

specifies the variable in the input data set to be used to decide whether an observation is used for one of the three roles: training, validation, or testing. You must specify at least one of the roles and its corresponding value. An observation is used for a role only if the value of the variable equals the value that you specify for that role. For example, if you specify TRAIN='value', then the observation is used for training only if the value of variable equals value. This variable is ignored in the holdout data table.

Last updated: August 06, 2026