Automated Fraud Detection Test Generation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Online communities face challenges in identifying and mitigating fraudulent users, who can cause harm and annoyance by sending unsolicited requests or scams, leading to potential financial and personal victimization.
Innovation Solution
A system comprising a fraudulent user detector and test generator that analyzes user data using equations and thresholds to identify fraudulent users, including the use of control and general population databases to adapt to changing user behaviors and reduce false positives.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If manual review of user profiles is performed to identify fraudulent users, then detection accuracy improves, but administrative workload and time increase
Solution Approach 1:
The system performs self-service by automatically generating and applying tests to user profiles without requiring manual administrative review. The automated test generation system creates detection rules based on fraudulent user characteristics and applies them consistently across all users, eliminating the need for manual intervention while maintaining high detection accuracy.
Solution Approach 2:
The patent replaces the mechanical manual review process with an automated computer-based test generation and application system. The system automatically analyzes user profiles against generated tests, substitutes human administrative work with algorithmic processing, and maintains consistent, objective fraud detection without human intervention.
2Reliability
If strict fraud detection tests are applied to all users, then fraudulent users are identified more effectively, but legitimate users may be incorrectly flagged (false positives)
Solution Approach 1:
The system dynamically adjusts test parameters and thresholds based on the specific characteristics of fraudulent user profiles being detected. By modifying test parameters to match the nuanced behaviors of different fraudulent user types, the system maintains high reliability in identification while reducing false positives through adaptive parameter adjustment rather than applying fixed strict rules.
Solution Approach 2:
The test generation system is dynamic, creating adaptable detection rules that can adjust to different user behaviors and contexts. The system evolves test parameters based on learned patterns from fraudulent user profiles, allowing flexible detection that maintains accuracy while minimizing incorrect flagging of legitimate users through contextual adaptation.
Data Source
AI summary
Member profile information for a control set of one or more control members and for a fraudulent set of one or more fraudulent members are obtained. Each member in the control set is at least believed to be legitimate and each member in the fraudulent set is at least suspected of being fraudulent. A test associated with identifying fraudulent members is generated using the member profile information for the control set and for the fraudulent set; the test inputs one or more pieces of member profile information for a member being tested.