> ML_DATASET // CREDIT-CARD-FRAUD-DETECTION_v1.0
Credit Card Fraud Detection Dataset (ULB MLG)
Université Libre de Bruxelles (MLG) · Banking & Financial Fraud · 284,807 transactions, 30 PCA features
Banking & Financial FraudOpen Data Commons Open Database License (ODbL)284,807 transactions, 30 PCA featuresopen
Dataset Profile & Characteristics
Label Type:Binary (0: Normal, 1: Fraudulent; 0.172% fraud rate)
Languages:en
License Tier:permissive-open-source
Modalities:tabular
Intended Use
- Imbalanced classification research
- Cost-sensitive anomaly detection benchmarking
Prohibited / Discouraged Use
- Modern card fraud rules without adaptation to chip/contactless
Bias, Leakage & Privacy Risk Analysis
Privacy / Sensitive Data Risks:
Features anonymized via PCA to protect cardholder identities.
Known Bias:
European cardholder transaction sample from September 2013.
Known Benchmark Leakage:
Time variable reflects two consecutive days; ordering must be strictly preserved.