When using gradient boosting in SAS Viya's gbtreetrain CAS action, what does the SUBSAMPLERATE= parameter control?
Select an answer to reveal the explanation.
Short Explanation
Here's the deal — you use SUBSAMPLERATE= to set the fraction of training observations randomly sampled without replacement to build each tree. This is stochastic gradient boosting, and it helps reduce variance and prevent overfitting. It doesn't control input variables, learning rate, or branch size.
Full Explanation
SUBSAMPLERATE= is row subsampling for stochastic gradient boosting: it sets the fraction of training observations randomly sampled without replacement for each tree, reducing variance and helping prevent overfitting. The input-variable sampling option describes column/feature subsampling, not SUBSAMPLERATE=. The learning-rate option describes shrinkage, not row sampling. The minimum-branch option describes minimum leaf or branch size.