Return Fraud Risk Scoring

ML scores return requests for fraud or abuse risk using transaction history, customer behavior, item category, timing, and known loss patterns.