The HPFOREST Procedure

PARTITION Statement

PARTITION ROLEVAR=variable <TRAIN='value'> <VALIDATE='value'> ;

The PARTITION statement specifies how to divide the input data set into a training subset and a validation subset. Variable names the variable in the input data set whose values are used to assign roles to each observation.

You can specify following suboptions:

TRAIN=value: requests that an observation be assigned to training if it matches value.
VALIDATE=value: requests that an observation be assigned to validation if it matches value.

If you do not specify the TRAIN= suboption, then any observation whose role is not determined by the VALIDATE= suboption is assigned to the training role.

If validation data is available, and the SCOREPROLE= PROC option is set to VALID, then predictions are based on the validation data in the leaves instead of the training data.

Last updated: July 02, 2020