Researchers have introduced FinFraudBench, a new benchmark designed to improve financial fraud detection using heterogeneous graphs. Existing benchmarks often oversimplify financial data into homogeneous graphs, failing to capture the complex, multi-entity relationships present in real-world systems. FinFraudBench addresses this by providing two large-scale datasets, CreditCard-Fraud and BankTrans-Fraud, which include six financial entity types and fourteen edge types, mimicking realistic conditions like extreme class imbalance and limited labels. The benchmark also establishes a standardized evaluation protocol to better assess the practical effectiveness of graph-based fraud detection methods. AI
IMPACT Aims to improve the practical effectiveness of graph-based methods for financial fraud detection by providing more realistic datasets and evaluation metrics.
RANK_REASON The item is a research paper introducing a new benchmark dataset and evaluation protocol for a specific AI task. [lever_c_demoted from research: ic=1 ai=1.0]
- arXiv
- BankTrans-Fraud
- Connected Papers
- CreditCard-Fraud
- FinFraudBench
- Hugging Face
- Litmaps
- scite Smart Citations
AI-generated summary · Google Gemini · from 1 sources. How we write summaries →
