How do random subspaces differ from ordinary bagging in what is sampled?
Classic MLLevel: IntermediateBagging sampling choices
How do random subspaces differ from ordinary bagging in what is sampled?
Hints
- There are two matrix axes.
- Replacement and sampling axis are independent choices.
Answer
Ordinary bagging samples observations, typically with replacement. Random subspaces sample feature subsets; combined schemes can sample both rows and features to diversify estimators.
Source and licence: scikit-learn 1.7.2 — ensemble · BSD-3-Clause · answer: Quantoros editorial