A data scientist in SAS Viya wants to apply oversampling to a minority class before training PROC HPFOREST. Which SAS Viya capability generates a synthetic oversampled training table using SMOTE (Synthetic Minority Oversampling Technique)?
Select an answer to reveal the explanation.
Short Explanation
SMOTE in SAS Viya isn't a PROC at all, it's a CAS action: smoteSample, in the smote action set. You call it from PROC CAS (or Python, Lua or R) and it builds brand-new minority rows by blending real neighbors, not photocopies. That's the difference between a real synthetic sample and just duplicating the rows you already have.
Full Explanation
SAS Viya provides SMOTE as a CAS action rather than a procedure: the Synthetic Minority Oversampling Technique action set, named smote, contains a single action, smoteSample. The action creates new minority-class observations by interpolating between an observation and its nearest minority-class neighbors, so the output table contains genuinely new synthetic rows. Because it is a CAS action, you run it from PROC CAS with CASL, or from the Python, Lua or R interfaces, and then train the downstream model on the resulting table. PROC HPCLUS with METHOD=SMOTE is wrong because HPCLUS performs k-means and k-modes clustering and has no SMOTE method. PROC SURVEYSELECT with METHOD=URS and REPS= is wrong because unrestricted random sampling with replacement only duplicates existing records, which adds no new information and encourages overfitting. PROC HPSAMPLE with OVERSAMPLING=SMOTE is wrong because HPSAMPLE supports simple random and stratified sampling and has no SMOTE option. An exam caveat is the word procedure: SMOTE lives in the action layer, so answers that force it into a PROC option are distractors. To confirm, load the action set with builtins.loadActionSet and compare class counts before and after smoteSample runs.