The HPFOREST Procedure

TREATMENT Statement

TREATMENT variable <ORDER=order>;

The TREATMENT statement names a binary treatment variable and estimates how much the prediction of an observation changes when the treatment changes. The analysis of differential treatment effects is sometimes called incremental response modeling (IRM). The target variable specified in the TARGET statement must have an interval or binary measurement level. For information about how PROC HPFOREST uses the treatment variable, see the section Modeling Incremental Response from a Treatment. Creation of forests with a treatment variable is experimental in PROC HPFOREST.

ORDER=order: specifies the sorting order of the treatment values. The smallest and largest sorted values are coded as 0 and 1 respectively in the names of variables in the output data set of the SCORE statement. Table 1 provides the recognized values of order.

Last updated: May 25, 2022