Why can mislabeled observations become especially influential in AdaBoost?
Classic MLLevel: IntermediateReweighted hard examples
Why can mislabeled observations become especially influential in AdaBoost?
Hints
- Follow weight changes after each mistake.
- Not every hard example is informative.
Answer
Misclassified examples receive increasing attention through reweighting. Persistent label errors can repeatedly attract later weak learners, so label quality and regularization matter.
Source and licence: scikit-learn 1.7.2 — ensemble · BSD-3-Clause · answer: Quantoros editorial