The HPSPLIT Procedure

OUTPUT Statement

  • OUTPUT OUT=SAS-data-set;

The OUTPUT statement creates a data set that contains one observation for each observation in the input data set. The OUT= data set contains the following:

  • the response variable

  • any variables that you specify by using the ID statement

  • the observation’s assigned leaf number

  • the observation’s assigned node number

You can use the leaf number in conjunction with the RULES statement to examine the tree in more detail. You can use the node number in conjunction with the table that is produced using the NODES option to investigate the tree.

In addition, for regression trees the OUT= data set contains the following:

  • the prediction for this observation

  • the average value within the observation’s assigned leaf in the validation partition, if you specify a validation partition in the PARTITION statement

In addition, for classification trees the OUT= data set contains the following for each response variable level:

  • the fraction of training partition weight within the assigned leaf for the observation

  • the fraction of validation partition weight within the assigned leaf for the observation, if you specify a validation partition in the PARTITION statement