Why is RepeatedKFold not a valid direct splitter for crossvalpredict?
Classic MLLevel: IntermediateOne prediction per row
Why is RepeatedKFold not a valid direct splitter for cross_val_predict?
Hints
- Check the output's one-row/one-prediction contract.
- Repetition changes prediction multiplicity.
Answer
cross_val_predict requires each sample to appear in a test fold exactly once. RepeatedKFold produces several held-out predictions per row, so aggregation must be implemented explicitly.
Source and licence: scikit-learn 1.7.2 — cross validation · BSD-3-Clause · answer: Quantoros editorial