3.2Standard
High-confidence predictions are automated.
Make confidence match actual correctness.
Choose two evaluations.
× Incorrect
Compare accuracy, errors, refusals, and subgroup differences by confidence band using separate calibration data, then set automation and human-review thresholds.
Detailed explanation
IncorrectShowing a score removes the need for evaluation
The score itself must be validated.
CorrectUse separate calibration data and define automation and human-review thresholds
Confidence is safer to operationalize.
IncorrectA score of 0.9 automatically means 90 percent accuracy
Calibration is not established.
CorrectCompare accuracy, errors, refusals, and subgroup differences by confidence band
The score's meaning is tested.
Check official documentation
Compare the following points with the official documentation.
AWS公式AIF-C01 Domain 3.2・3.4の信頼度、評価、人的レビューを確認する。
Official references
What to confirm
モデルスコアを実際の正しさと比較し、処理の境界へ利用できる。
Key points
- Confidence
- Calibration
- Threshold
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
Confidence calibration
Aligning predicted confidence values with the observed frequency of correct results.
Review threshold
A boundary at which a case is routed to a person instead of automated handling.
Question IDAWS-AIF-399
EnvironmentAWS公式AIF-C01試験ガイドとAWS公式ドキュメントの確認
Last technical review2026-08-20
Report an error or rights issue