Data Quality Assurance in Credit Card Fraud Detection
summary
Claims that 2023 fraud causes about $10B in US losses and that fraud makes up only 0.5–1% of transactions drive a focus on data quality. Practical fixes include replacing free‑text dispute reasons with dropdown validation to stop wrong labels, blocking 9999 placeholder values detected by a rule (e.g. >10% values >2σ) sent to alerts, and linking split identities via entity resolution into graph detections; tools cited: data lakehouse, data catalog, monitoring and decision engines.
AI-written summary · about
chapters (10)
- 0:00Data quality in fraud detection
- 3:58Why data quality matters
- 8:36Key data points for fraud
- 13:15Common data quality issues
- 16:53Data quality tools
- 20:08Fixing wrong labels
- 23:46Detecting abnormal values
- 29:15Entity resolution for customers
- 34:20Challenges and solutions
- 35:49Future of AI and data quality
related talks
35mBig Data in the Fight Against Fraud
47mTrain, Evaluate, Repeat: Building a Credit Card Fraud Detection System
14mNext-Gen Fraud Strategies Use Data to Onboard Customers Safely
13mHarnessing Machine Learning to Combat E-commerce Fraud
26mAI-Powered Fraud Detection in Cloud-Native Digital Payments
24mTop Fraud Trends in Crypto for 2024 and How to Stay Ahead in 2025
08mAI-Powered Fraud Detection in Financial System
35mAI Powered Fraud Detection for Payments & Transactions