Sub-network Risk Scoring for Financial Fraud Detection
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current methods for detecting and investigating complex fraud schemes in financial networks are inefficient, particularly in large networks, as existing predictive algorithms struggle to effectively identify high-risk entities and sub-networks for suspicious financial activity.
Innovation Solution
A system and method that involve receiving network data, identifying seed entities based on predefined rules, generating sub-networks, updating risk scores, and calculating risk scores for these sub-networks, allowing for iterative processing and application of predictive algorithms to efficiently detect suspicious financial activity.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If existing predictive algorithms are applied to large financial networks, then risk scoring capability is provided, but the algorithms become difficult to apply and efficiency deteriorates
Solution Approach 1:
The patent divides the large financial network into multiple sub-networks based on entity relationships and risk characteristics. This segmentation allows predictive algorithms to be applied to smaller, more manageable sub-networks rather than the entire large network, improving computational efficiency while maintaining risk scoring capability through iterative processing of segments.
Solution Approach 2:
The patent introduces a hierarchical dimension by creating multiple levels of sub-networks (first-level, second-level, etc.) based on risk scores and relationship depths. This dimensional transformation allows the system to process network data at different granularities, making large-scale network analysis tractable while preserving comprehensive risk assessment capabilities.
2Reliability
If the entire network is examined for suspicious activity, then complete coverage is achieved, but computational efficiency deteriorates in large networks
Solution Approach 1:
The network is segmented into sub-networks based on entity relationships and risk characteristics, allowing focused analysis of relevant portions rather than exhaustive examination of the entire network. This maintains detection coverage for high-risk areas while improving efficiency by excluding low-risk segments from intensive analysis.
Solution Approach 2:
The system performs preliminary risk scoring and identifies high-risk entities before conducting detailed suspicious activity analysis. This preliminary action filters the network data to focus computational resources on entities and sub-networks most likely to contain suspicious activity, achieving efficient targeted detection.
3Measurement precision
If deeper understanding of relationships between entities is obtained, then fraud detection accuracy is improved, but system complexity increases
Solution Approach 1:
The patent analyzes relationships at multiple hierarchical levels by creating sub-networks at different depths (first-level sub-networks, second-level sub-networks, etc.). This dimensional approach enables deeper understanding of entity relationships through iterative expansion while managing complexity through structured, level-by-level processing rather than attempting to analyze all relationships simultaneously.
Solution Approach 2:
The system dynamically adjusts the depth and scope of relationship analysis based on risk scores and detected patterns. High-risk entities trigger deeper multi-level sub-network analysis, while low-risk entities receive simpler analysis, making the system adaptively complex rather than uniformly complex throughout.
Data Source
AI summary
A system or method may include receiving, by a processor, data describing a network, wherein the network includes a plurality of entities and links describing relationships between the plurality of entities. The method may further include identifying a set of seed entities from the plurality of entities based on predefined rules. The method may further include generating a set of sub-networks based on the set of identified seed entities, wherein each of the sub-networks may include one or more other entities of the plurality of entities having at least one link to the at least one seed entity. The method may further include calculating a risk score for each of the generated sub-networks.