BookmarkSubscribeRSS Feed

HPSplit is in Base SAS/STAT as a standard offering. Random forests are a staple of predictive modeling. Move HP Forest out of SAS Enterprise Miner and put it in the SAS/STAT product. Random forests are 20 years old now, it is nowhere near a novel technique.

 

There are other packages that need moving over but this one is a no-brainer, at least to me.

6 Comments
nohadra
Calcite | Level 5

Yes please! Support vector machines and ensemble tree models (random forests and boosted trees) are algorithms dating back to the 90s. I concur it's time to move them to SAS/STAT.

Thanks for your excellent support 

Noah_
Calcite | Level 5

Just saw this, I definitely agree with both users. It's about time to include HPFOREST/HPSVM and other basic Enterprise Miner procedures in SAS/STAT. These algorithms aren't novel anymore...

nohadra
Calcite | Level 5

I totaly agree. any updates on this ? it`s been almost 2 years since it was submitted on february 1st 2022... 

jleirer
SAS Employee
Status changed to: Suggestion Closed

Decision trees currently ships with SAS/STAT as PROC HPSPLIT.  Random Forest is currently provided in SAS Viya Machine Learning in the FOREST Procedure.  

jleirer
SAS Employee

I also want to note that SVM is available in the SVMACHINE Procedure in SAS Viya Machine Learning.  (https://go.documentation.sas.com/doc/en/pgmsascdc/v_053/casml/casml_svmachine_overview.htm)

Noah_
Calcite | Level 5

Thank you @jleirer unfortunately the licence for SAS VIYA is more onerous than SAS/STAT. Have a nice day.