The HPREDUCE Procedure

PROC HPREDUCE Features

The HPREDUCE procedure conducts a variance analysis and reduces dimensionality by selecting the variables that contribute the most to the overall variance of the data (or the dependent variables). The following list summarizes the basic features of the HPREDUCE procedure:

Variable selection is based on covariance analysis.
Analysis can be performed on a massively parallel SAS high-performance appliance.
Input data can be read in parallel when the data source is the appliance database.
Computation of the CORR, COV, or SSCP matrix is distributed.
Computation of the variable selection steps is distributed.
All phases of analytic execution use of high degree of multithreading.
Both supervised and unsupervised variable selection are supported.
Multiple response variables are supported in variable selection for regression.
The CLASS statement supports categorical inputs.
The REDUCE statement supports main and interaction effects.
The OUTCP statement supports outputting a CORR, COV, or SSCP matrix.