The HPSPLIT Procedure

OUTPUT Statement

OUTPUT <output-options> </ subtreestat-option> ;

The OUTPUT statement allows several SAS data sets to be created.

You can specify the following output-options:

GROWTHSUBTREE=SAS-data-set

writes to the specified SAS-data-set a table that contains the requested statistical metrics of the subtrees that are created during growth.

IMPORTANCE=SAS-data-set

writes the importance of each variable to the specified SAS-data-set.

NODESTATS=SAS-data-set

writes a description of the final tree to the specified SAS-data-set.

PRUNESUBTREE=SAS-data-set

writes to the specified SAS-data-set a table that contains the requested statistical metrics of the subtrees that are created during pruning.

You can specify the following subtreestat-option:

SUBTREESTATS=(metric <metric ...>)

specifies the statistical metrics to write to the subtree data sets. The iteration number, number of leaves, and tree number are always provided.

You can specify one or more of the following metrics.

ENTROPY

calculates the entropy of the subtree.

GINI

calculates the Gini statistic of the subtree.

ASE

calculates the average square error of the subtree.

MISC

calculates the misclassification rate of the subtree.

SSE

calculates the sum of squares error of the subtree.

ALL

enables all the statistics.