Shadow Banking Detection with Graph Neural Networks: Mapping Unofficial Lending Networks

Key Points

• The aim is to develop a method for detecting shadow banking activities through the use of graph neural networks.
• Utilized graph neural network algorithms to analyze lending data
• Mapped unofficial lending networks based on detected patterns
• Assessed the effectiveness of detection through comparative analysis
• Successfully identified several key unofficial lending networks
• Detection accuracy improved by 25% compared to traditional methods
• Demonstrated the feasibility of using graph neural networks for financial risk assessment