Multi-Account Payment Fraud Analytics via Segmented Profiling
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Traditional fraud detection systems are inadequate for monitoring transactions involving multiple funding accounts, as they are designed for single-account scenarios and fail to account for dynamic funding choices and patterns across accounts, leading to challenges in detecting fraudulent activities.
Innovation Solution
A computer-implemented fraud analytic system and method that generates a payment instrument profile characterizing past and current activity across multiple funding accounts, using variables to detect abnormalities and generate a fraud score, incorporating techniques like Bayesian statistics, self-calibrating outlier analytics, favorite lists, and cross-funding account dynamics to monitor and manage funding account profile segments effectively.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If traditional single-account fraud detection systems are used, then the system complexity remains low, but the system cannot detect fraudulent activities involving multiple funding accounts
Solution Approach 1:
The patent segments the fraud detection system into multiple independent analytical modules: transaction pattern analysis, funding account selection analysis, behavioral biometrics, and anomaly detection. Each module processes specific aspects of multi-account transactions independently, then integrates results to produce comprehensive fraud scores. This segmentation enables the system to handle complex multi-account scenarios while maintaining manageable system architecture.
Solution Approach 2:
The patent creates a universal fraud detection framework that handles both single-account and multi-account scenarios through a unified analytical engine. The system uses common data structures and processing logic that automatically adapt to the number of funding accounts involved, eliminating the need for separate detection systems for different account configurations.
2Measurement precision
If traditional fraud detection methods are applied to multi-funding accounts, then implementation remains simple, but the detection accuracy and visibility into customer behavior deteriorates
Solution Approach 1:
The patent adds a new dimension to fraud detection by incorporating funding account selection patterns as an additional analytical layer. Instead of only analyzing transaction characteristics, the system evaluates which funding accounts are selected, how frequently, and under what conditions. This dimensional expansion provides deeper insights into customer behavior and improves fraud detection accuracy by capturing patterns that span multiple accounts.
Solution Approach 2:
The patent implements feedback mechanisms where detection results and anomaly patterns are continuously fed back into the analytical model to refine future detections. The system learns from observed funding account selection patterns and transaction behaviors, adjusting thresholds and parameters dynamically to improve detection accuracy while adapting to legitimate changes in customer behavior.
Data Source
AI summary
A system and method for multiple funding account payment instrument analytics is disclosed. A payment instrument profile is generated for a payment instrument that characterizes past activity on the payment instrument and past activity across one or more funding accounts associated with the payment instrument. Current activity on the payment instrument and current activity across one or more funding accounts associated with the payment instrument are monitored to detect an abnormality against the payment instrument profile. A fraud score for the current activity on the payment instrument and current activity across one or more funding accounts associated with the payment instrument is generated. The fraud score is based on a quantified extent of the abnormality.