Fraud Probability Engine for Hybrid Payment Networks
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current fraud control methods in payment networks face challenges in accurately determining the likelihood of fraudulent data transfers, especially in account-to-account transactions where historical data is limited, leading to a lack of interoperability between payment card and account-based systems.
Innovation Solution
A computer-implemented method assesses the probability of fraudulent activity by determining associations between payment credentials, accessing historical data from both card settlement and merchant accounts, and using a fraud probability engine to generate a fraud probability value based on past transactions, even when historical data for specific account-to-account transactions is scarce.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If fraud control methods use historical data from specific account-to-account transactions to assess fraud probability, then measurement precision improves, but reliability deteriorates when historical data is limited or unavailable
Solution Approach 1:
The patent combines historical transaction data from multiple sources including card-based transactions and account-based transactions into a unified fraud assessment model. This merging allows the system to leverage extensive historical data from card transactions to improve fraud detection reliability for account-to-account transactions, even when specific historical data for the latter is limited.
Solution Approach 2:
The fraud probability engine is designed to work with multiple types of transaction data (both card-based and account-based) using a universal assessment methodology. This multi-functionality enables the system to apply the same fraud detection logic across different transaction types, allowing historical data from one type to inform assessments of another type.
2Device complexity
If separate networks are used for payment card transactions and account-to-account transactions, then device complexity is reduced, but interoperability deteriorates
Solution Approach 1:
The patent introduces a fraud probability engine as an intermediary component that receives transaction data from both card-based networks and account-based networks. This mediator translates and harmonizes data from different network protocols and formats into a common assessment framework, enabling interoperability while maintaining the independence of each network.
Solution Approach 2:
The system adds a new dimension to fraud detection by incorporating cross-network historical data. Instead of analyzing transactions within a single network dimension, the system integrates data from multiple network dimensions (card-based and account-based), creating a multi-dimensional fraud assessment capability that enhances interoperability.
Data Source
AI summary
A computer implemented method of assessing the probability of fraudulent activity in an account to account payment is provided. The method comprises determining, based on previous credit balance settlement transactions between a first account and a card settlement account, that a first set of payment credentials is associated with the first account. It is determined that a second set of payment credentials are associated with the merchant account. A database comprising historical data indicating details of past payment transactions between the holder of the first set of payment credentials to the holder of the second set of payment credentials is accessed. A fraud probability engine is used to generate, based on the historical data, a fraud probability value for a transaction from a first account to a second account, wherein the fraud probability value is an estimate of the probability of the transaction being fraudulent.