Why distinguish detecting outliers in a contaminated training set from detecting novelty relative to a clean reference…
Classic MLLevel: IntermediateReference data assumption
Why distinguish detecting outliers in a contaminated training set from detecting novelty relative to a clean reference set?
Hints
- Compare what training data is assumed to contain.
- Ask whether scored rows were used to fit.
Answer
The first must identify unusual members of the data used to fit; the second learns normality and scores new observations. Estimator APIs and contamination assumptions can differ materially.
Source and licence: scikit-learn 1.7.2 — outlier detection · BSD-3-Clause · answer: Quantoros editorial