PayFlow processes roughly 8 million card transactions per day across web and mobile checkout. The risk team wants a fraud detection model that scores each transaction before authorization so high-risk payments can be declined or routed to manual review.
You are given a historical transaction dataset labeled after chargeback resolution.
| Feature Group | Count | Examples |
|---|---|---|
| Transaction attributes | 12 | amount, currency, merchant_category, entry_mode, is_international |
| Customer behavior | 10 | transactions_last_1h, avg_amount_30d, device_count_7d, failed_attempts_24h |
| Device & network | 8 | device_id_hash, ip_country, proxy_flag, browser_family |
| Merchant signals | 6 | merchant_risk_score, merchant_chargeback_rate, merchant_country |
| Temporal features | 5 | hour_of_day, day_of_week, days_since_account_open |
A strong solution should achieve high recall on fraud while keeping false positives low enough that approval rate is not materially harmed. Target recall >= 75% at precision >= 20%, PR-AUC >= 0.30, and p95 online inference latency < 50 ms.