Wykrywanie oszustw
1 1 related system
Identification of transactions, applications, or behavior exhibiting signs of fraud in real-time or batch. AI/ML models leverage tabular data (transaction features, device, geolocation, account history) for binary legit/fraud classification and risk scoring. Main domains: card fraud, account takeover, AML/KYC, insurance fraud. Tabular foundation models (TabPFN) and gradient boosting (XGBoost, LightGBM) remain the industry standard.
