A classifier scores 99% on training data but 65% on unseen test data.
Investigate the difference before deployment.
Which description is most appropriate?
The gap is a sign of possible overfitting: the model fits training examples but generalizes poorly.
Detailed explanation
High training performance with much lower test performance indicates poor generalization.
The cause and generalization must be investigated first.
This destroys the independent evaluation.
The gap alone does not prove whether leakage exists.
Check official documentation
Compare the following points with the official documentation.
AWS公式AIF-C01ガイドDomain 1.3の過学習・汎化・評価データを確認する。
Official references
What to confirm
訓練とテスト性能の差を過学習の兆候として説明できる。
Key points
- Overfitting
- Generalization
- Test data
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
Held-out evidence
A test result is informative only when the data did not guide parameter fitting or repeated tuning.
Production readiness
High training accuracy alone is not evidence that a model will work on new cases.