A hiring model still shows regional disparities after removing location.
Investigate fairness rather than assuming the problem is solved.
Choose two appropriate investigations.
Check whether remaining features proxy for protected attributes and monitor group-specific errors and impacts.
Detailed explanation
Averages can hide group harm.
Proxy variables and sampling or label bias may remain.
The operational consequence may be hidden by overall averages.
Postal codes or behavior may encode the removed attribute indirectly.
Check official documentation
Compare the following points with the official documentation.
AWS公式AIF-C01ガイドDomain 4.1のバイアス、公平性、データ品質を確認する。
Official references
What to confirm
代理変数とグループ別影響を確認する理由を説明できる。
Key points
- Proxy
- Group metrics
- Impact
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
Proxy variable
An input correlated with a protected attribute that can reproduce its influence indirectly.
Group evaluation
Measuring quality, errors, and outcomes separately for defined populations.