Dynamic Merchant Graph for Fraud Detection
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional fraud detection systems fail to effectively capture the complex and rapidly changing social network-like behavior of purchaser-merchant interactions, leading to inadequate detection of fraudulent transactions, as they often rely on static rules that do not account for individual differences in shopping behavior and miss fast-moving changes in transaction patterns.
Innovation Solution
A fraud detection system that utilizes a dynamically evolving merchant relationship graph to assess the likelihood of fraudulent transactions by analyzing transaction patterns and relationships between financial accounts and merchants, incorporating a transaction validator that computes a merchant relatedness score based on historical transaction data and relationship tables to determine the validity of transactions in real-time.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If conventional static rules are used for fraud detection, then the system is simple to implement, but it fails to capture complex and rapidly changing social network behavior patterns
Solution Approach 1:
The patent applies the dynamics principle by transitioning from static fraud detection rules to a dynamic graph-based system that continuously evolves with transaction data. The merchant relationship graph is updated in real-time as new transactions occur, allowing the system to adapt to changing shopping behaviors and fraud patterns dynamically rather than relying on predetermined static rules
Solution Approach 2:
The patent implements dimensionality change by introducing a graph-based structure that adds relational dimensions to traditional transaction analysis. Instead of analyzing transactions in isolation or using simple rule-based filters, the system constructs a multi-dimensional merchant relationship graph that captures complex social network patterns, peer relationships, and transactional connections across multiple layers
2Measurement precision
If global rules are applied to all shoppers, then the system is easier to manage, but it misses individual differences in shopping behavior and fast-moving changes in patterns
Solution Approach 1:
The patent applies local quality by creating individualized fraud detection profiles for each shopper based on their unique transaction history and relationships in the merchant graph. Instead of applying uniform global rules, the system generates personalized risk assessments that reflect each user's specific shopping behavior patterns, peer networks, and transactional characteristics
Solution Approach 2:
The system implements feedback mechanisms where transaction outcomes and fraud detection results continuously update the merchant relationship graph and individual user profiles. This feedback loop allows the system to learn from new data, refine detection accuracy, and adapt to emerging fraud patterns while maintaining individualized assessments for each shopper
3Ease of operation
If simple pass-fail tests like passwords or biometrics are used, then the system is easy to operate, but legitimate customers and vendors find them too troublesome
Solution Approach 1:
The patent implements self-service by enabling the system to automatically perform fraud detection and risk assessment without requiring users to manually provide additional verification information. The merchant relationship graph continuously analyzes transaction patterns in the background, automatically flagging suspicious activities while allowing legitimate transactions to proceed seamlessly without user intervention
4Adaptability or versatility
If maximum monetary limits are set, then the system provides simple control, but it allows many small fraudulent transactions while blocking large legitimate transactions
Solution Approach 1:
The patent applies parameter changes by dynamically adjusting fraud detection thresholds and risk assessment parameters based on the evolving merchant relationship graph and individual user profiles. Instead of using fixed monetary limits, the system modifies detection sensitivity, risk weights, and approval thresholds in real-time based on transaction patterns, user history, and relationship data
Data Source
AI summary
Systems and methods for enhanced detection of fraudulent electronic transactions are disclosed. In one embodiment, a system uses the ongoing stream of transactions to construct and maintain a dynamically evolving merchant relationship graph. When a proposed transaction is submitted to the system, the system computes a predicted likelihood that the given account would make a transaction with these characteristics with the given merchant. The graph is used to compute transitive relatedness between merchants which may be indirectly associated with one another, as well as to compute aggregate relatedness, when there are multiple avenues of relationship between two merchants.