Transaction Pathway Analysis for Fraud Pattern Detection
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current fraud detection systems in payment transactions face challenges in real-time identification of fraud patterns and adapting to evolving fraud tactics, particularly with the EMV migration in the United States, which leads to changes in fraud types and regions, resulting in increased monetary losses and the need for continuous model updates.
Innovation Solution
The method involves building a decision tree based on a training dataset, recording pathway transversal information, calculating mean and variance of class probabilities, and determining deviation metrics between reference and new transaction datasets to identify pattern changes and feature contributors, generating alerts and reports based on predetermined criteria.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If traditional fraud detection models are used, then fraud detection capability is maintained, but the system cannot adapt to evolving fraud tactics and pattern changes
Solution Approach 1:
The system performs preliminary actions by continuously monitoring transaction pathways and detecting pattern changes before they become significant threats. The pathway analysis framework proactively identifies deviations from normal transaction patterns, allowing the system to adapt fraud detection rules in advance of major fraud tactic shifts, thereby maintaining reliability while improving adaptability.
Solution Approach 2:
The system implements feedback mechanisms by continuously comparing actual transaction pathways against expected patterns and using the detected deviations to refine fraud detection models. The pathway analysis provides feedback on emerging fraud tactics, enabling the system to adapt and improve its detection capability over time while maintaining reliable operation through continuous learning.
2Measurement precision
If continuous model updates are performed to detect new fraud patterns, then fraud detection accuracy improves, but system complexity and computational resources increase
Solution Approach 1:
The system segments the fraud detection process into distinct pathway analysis components, where transactions are evaluated along specific decision tree pathways rather than using monolithic models. This segmentation allows for targeted updates of specific pathway rules based on detected pattern changes, improving detection accuracy while reducing overall system complexity by updating only necessary segments.
Solution Approach 2:
The system applies partial updates to fraud detection models by modifying only the specific pathways and rules that show significant deviations, rather than performing complete model retraining. This partial action approach maintains high detection accuracy for affected pathways while reducing computational resources and system complexity compared to full model updates.
3Loss of information
If detailed pathway analysis is performed on all transactions, then fraud pattern detection capability improves, but processing time and computational resources increase
Solution Approach 1:
The system applies local quality by performing detailed pathway analysis only where necessary - specifically at decision nodes and pathways showing deviations from expected patterns. Normal transactions follow standard pathways with minimal analysis, while suspicious transactions trigger detailed pathway examination. This approach preserves critical fraud pattern information while maintaining real-time processing capability.
Solution Approach 2:
The system performs partial pathway analysis by evaluating only the relevant decision nodes and pathways for each transaction rather than analyzing all possible pathways. The analysis depth is adjusted based on transaction risk indicators, applying excessive detail only when needed for suspicious patterns. This maintains comprehensive fraud detection capability while reducing average processing time for the overall transaction volume.
Data Source
AI summary
A method for detecting fraud and non-fraud pattern changes can be based on transaction pathway transversal analysis. A decision tree can be built based on a training dataset from a reference dataset. Pathway transversal information can be recorded along each pathway for the reference dataset. A first mean and a first variance of a class probability can be calculated of all samples over each pathway. A pathway distribution for a new transaction dataset under investigation and a second mean and a second variance of all samples of the new transaction dataset can be obtained. The second mean and the second variance can represent a fraud probability. The deviation metrics between one or more feature statistics of a feature along each pathway for the reference dataset and the new dataset can be determined on a local level. Feature contributors to pattern changes can be determined by analyzing the deviation metrics.