Compare models with a small dataset.
Prevent preprocessing and duplicate records from contaminating evaluation.
Choose two appropriate practices.
Fit preprocessing within each training fold and keep the same entity or duplicate records from crossing train and validation folds.
Detailed explanation
Overlap can overstate unseen-data performance.
Reproducibility and trust would be lost.
Validation information must not determine training preprocessing.
Global statistics can leak validation information.
Check official documentation
Compare the following points with the official documentation.
AWS公式AIF-C01ガイドDomain 1.3のデータ分割、前処理、リーケージを確認する。
Official references
What to confirm
交差検証で前処理と重複を管理する理由を説明できる。
Key points
- Cross-validation
- Preprocessing
- Duplicates
Notes
- Environment: AWS公式AIF-C01試験ガイドとAWS公式ドキュメントの確認
- Compare the definition with the AIF-C01 objective and the service documentation before applying it.
- Keep the concept separate from a particular answer choice or implementation step.
Foundation review
Cross-validation
Repeatedly splitting data into training and validation folds to estimate model performance.
Data leakage
Information from evaluation or the future influencing training and making performance look too good.