Dual-Process Fraud Detection for Lower False Positives
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing fraud detection methods in systems like financial, banking, and industrial systems suffer from high false-positive rates and fail to reliably detect all fraudulent activities, impacting system performance and customer experience.
Innovation Solution
A method and system that employs parallel iterative processes for fraud detection, using supervised and unsupervised learning to estimate legitimacy and illegitimacy scores based on data streams, with pre-processing to identify critical events and neural networks to analyze operator behavior, ensuring robust and accurate fraud identification.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If conventional fraud detection methods are used, then fraud detection capability is provided, but false-positive rates become too high
Solution Approach 1:
The fraud detection process is segmented into two independent parallel processes: a legitimacy process that calculates legitimacy scores based on legitimate operator patterns, and an illegitimacy process that calculates illegitimacy scores based on fraudulent operator patterns. Each process independently evaluates the same set of events but with different learned parameters, allowing the system to compare both perspectives and reduce false positives by requiring concordance between legitimacy and illegitimacy assessments.
Solution Approach 2:
The system changes parameters by maintaining separate parameter sets for legitimacy and illegitimacy processes. The legitimacy process learns parameters from databases of legitimate operator behaviors, while the illegitimacy process learns parameters from databases of fraudulent operator behaviors. This parameter differentiation enables the system to evaluate operator behavior through two distinct analytical lenses, improving detection accuracy.
2Reliability
If conventional fraud detection methods are used, then fraud detection is performed, but false-negative detection increases
Solution Approach 1:
The system implements feedback mechanisms where the legitimacy process and illegitimacy process continuously compare their score assessments. When there is a significant discrepancy between legitimacy scores and illegitimacy scores for the same operator events, the system triggers further analysis or flags for manual review. This feedback loop ensures that potential fraud cases are not missed and allows the system to learn from detection outcomes to improve future detection accuracy.
3Measurement precision
If dual parallel processes are implemented, then detection accuracy improves, but system complexity increases
Solution Approach 1:
The system uses a universal event processing framework that handles both legitimacy and illegitimacy assessments through the same computational infrastructure. Both processes use identical data structures, event representations, and processing pipelines, differing only in their learned parameters. This universal approach reduces system complexity by avoiding duplication of processing logic while maintaining the benefits of dual-process analysis.
Data Source
AI summary
The invention relates to a method and a system for the automated detection of the risk of fraud in a monitored system, based on data streams generated by said monitored system and characterizing events performed or generated by operators in said monitored system. The method includes: a pre-processing (30-38) of at least one set of data recorded over a period of time, so as to obtain a subset of critical events associated with an operator; the iterative application of a first parameterized estimation process (52,56) for a risk of fraud, so as to obtain a first legitimacy score and a first associated probability of occurrence; the iterative application of a second parameterized estimation process (54,58) for a risk of fraud, so as to obtain a second legitimacy score and a second associated probability of occurrence, and the comparison (60) of the results of said first and second processes, for determining (64) whether said operator is a legitimate operator or a fraudulent operator.