Why train a stacking meta-model on out-of-fold base predictions rather than predictions from base models fitted on the…
Classic MLLevel: AdvancedMeta-model training leakage
Why train a stacking meta-model on out-of-fold base predictions rather than predictions from base models fitted on the same rows?
Hints
- The meta-model must see realistic base-model errors.
- Each row must be unseen by its generating base fit.
Answer
This answer is part of mentoring
Advanced questions come with a full breakdown and a mock interview with a working quant researcher. The first call is free.
Source and licence: scikit-learn 1.7.2 — ensemble · BSD-3-Clause · answer: Quantoros editorial