Dynamic Peer Grouping for Fraud Rule Optimization

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

Conventional transaction fraud detection systems face challenges in optimizing fraud rules, leading to either excessive rejection of legitimate transactions (false positives) or allowing fraudulent transactions due to inadequate capture rates, with entities unsure if their rule sets are optimal.

Innovation Solution

A system that analyzes transaction data from various entities, segments it using filters, identifies peer entities for comparison, and automatically generates and implements rules to adjust fraud rates, ensuring they align with peer performance metrics, thereby optimizing fraud detection.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Reliability

If stringent fraud rules are applied, then fraudulent transactions are blocked, but legitimate transactions are rejected (false positives)

Engineering Contradiction:

Improvefraud detection accuracyVSAvoidtransaction approval rate

Core Design Contradiction:

ReliabilityVSProductivity

Solution Approach 1:

The patent segments transactions into different groups based on characteristics such as transaction type, amount, merchant category, and geographic location. By applying different fraud rules to different segments rather than a uniform rule set, the system can maintain high fraud detection accuracy for high-risk segments while minimizing false positives in low-risk segments, thus resolving the contradiction between reliability and productivity

Inventive Principle:

Principle #1Segmentation

Solution Approach 2:

The system dynamically adjusts fraud rules and thresholds based on real-time analysis of transaction patterns, peer performance metrics, and identified opportunities. Rules are not static but adapt to changing conditions, allowing the system to maintain optimal fraud detection while minimizing unnecessary rejections of legitimate transactions

Inventive Principle:

Principle #15Dynamics

2Reliability

If fraud rules capture many transactions, then fraudulent transactions are blocked, but legitimate transactions are also denied

Engineering Contradiction:

Improvefraud capture rateVSAvoidtransaction throughput

Core Design Contradiction:

ReliabilityVSProductivity

Solution Approach 1:

Different fraud detection stringency levels are applied to different transaction types, merchants, and geographic regions based on their specific risk profiles. High-risk transactions receive more stringent scrutiny while low-risk transactions are processed with minimal friction, optimizing both fraud capture rate and transaction throughput by matching detection intensity to local risk characteristics

Inventive Principle:

Principle #3Local quality

3Measurement precision

If manual rule optimization is performed, then fraud detection accuracy may improve, but time and resources are consumed

Engineering Contradiction:

Improvefraud rule performance assessmentVSAvoidrule optimization time

Core Design Contradiction:

Measurement precisionVSLoss of time

Solution Approach 1:

The system automatically identifies performance gaps by comparing an entity's fraud metrics against peer benchmarks, generates optimized rule recommendations, and implements rules without requiring manual analysis. This self-service approach maintains high measurement precision for fraud rule performance while eliminating the time and resource costs of manual optimization

Inventive Principle:

Principle #25Self-service

Solution Approach 2:

The system continuously monitors fraud detection performance metrics, compares them against peer entities, and uses this feedback to automatically generate and adjust fraud rules. This closed-loop feedback mechanism enables continuous optimization of fraud detection accuracy without manual intervention, resolving the contradiction between measurement precision and time consumption

Inventive Principle:

Principle #23Feedback

4Ease of operation

If static peer groups are used for comparison, then assessment simplicity is maintained, but accuracy decreases due to irrelevant comparisons

Engineering Contradiction:

Improvepeer comparison simplicityVSAvoidperformance assessment accuracy

Core Design Contradiction:

Ease of operationVSMeasurement precision

Solution Approach 1:

Peer groups are dynamically reconfigured based on transaction characteristics, risk profiles, and performance metrics rather than using static predetermined groups. This allows the system to maintain ease of operation through automated peer selection while significantly improving measurement precision by comparing entities against truly relevant peers with similar transaction patterns and risk characteristics

Inventive Principle:

Principle #15Dynamics

Data Source

PatentUS10755280B2Segmented data analysis using dynamic peer groupings and automated rule implementation platform

Publication Date: 2020.08.25 VISA INTERNATIONAL SERVICE ASSOCIATION

AI summary

Described herein are systems and methods for providing accurate assessments of existing fraud rules while controlling for various transaction risks. In some embodiments, transaction data may be obtained from a number of entities and may be segmented by applying various filters. Once segmented, each segment may be analyzed to obtain metrics for a target entity. A separate set of peers is dynamically determined for each segment. The metric values for the target entity may then be compared to the metric values for the peers in the peer set to assess the target entity's performance with respect to each segment. Based on a variance of the target entity from its peers, the system may identify the segment as an opportunity. A rule may then be generated automatically based on the identified segment. In some embodiments, the rule may be added to a rule file for automatic implementation by the target entity.