ABC Bank Credit Card Approval Model: Balancing Accuracy, Fairness, and Explainability
Автор: Ryan
Загружено: 2026-03-12
Просмотров: 2
Описание:
Key highlights of this presentation include:
• Metric Selection: Optimizing for ROC-AUC and F1-Score to balance default risk with market share.
• Model Stability: Utilizing SMOTE and tuned LightGBM to overcome severe data imbalance and prevent overfitting.
• Responsible AI (RAI): Auditing the model for demographic fairness and mitigating bias.
• Explainability: Opening the "black box" using SHAP values for both Global (system-wide logic) and Local (individual applicant) transparency.
This presentation was created for the M5 Final Assignment in Practical Machine Learning.
Повторяем попытку...
Доступные форматы для скачивания:
Скачать видео
-
Информация по загрузке: