AML pipeline for training an XGBoost model on financial data using graph features.
-
Updated
Jul 6, 2026 - Python
AML pipeline for training an XGBoost model on financial data using graph features.
Alert-budget-aware evaluation methodology for AML transaction monitoring, on IBM's synthetic AMLworld benchmark. Account-day alert units, random-ranker nulls, attainable ceilings, and a publication gate that refuses any number no archived artifact supports.
To associate your repository with the amlworld topic, visit your repo's landing page and select "manage topics."