Impact-Based Fraud Detection for Crowd-Sourced Data
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Detecting fraudulent user activity in crowd-sourced submissions is challenging due to the difficulty in identifying and blocking organized groups of users who collectively influence ratings and traffic patterns.
Innovation Solution
A computing device is configured to perform impact-based fraud detection by identifying user groups and analyzing their submissions to prioritize and block high-impact groups, using surrogate identifiers to anonymize data and determine credibility scores based on submission impact.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If traditional fraud detection methods are used to identify fraudulent users in crowd-sourced submissions, then individual fraudulent submissions can be detected, but organized groups of users (fraud rings) having significant collective impact cannot be efficiently identified
Solution Approach 1:
The patent segments the analysis from individual user level to group level by identifying user groups based on shared characteristics (device identifiers, submission patterns, temporal proximity). This segmentation enables detection of organized fraud rings while maintaining individual user analysis capabilities, resolving the contradiction between detecting individual fraud and organized group fraud.
Solution Approach 2:
The patent creates surrogate identifiers that copy and anonymize user identity information while preserving grouping characteristics. These surrogate identifiers enable analysis of user groups without exposing individual identities, allowing efficient group-level detection while maintaining privacy and scalability.
2Reliability
If all user groups are analyzed equally for fraudulent activity, then comprehensive coverage is achieved, but processing resources are wasted on low-impact groups
Solution Approach 1:
The patent applies local quality by differentiating analysis depth based on group impact level. High-impact groups undergo comprehensive fraud analysis while low-impact groups receive minimal or no analysis. This targeted approach maintains reliable detection coverage for significant threats while reducing processing resource consumption on negligible cases.
Solution Approach 2:
The patent changes the parameter of analysis intensity based on group impact metrics. By calculating impact scores based on submission volume, target influence, and group characteristics, the system dynamically adjusts processing resources allocated to each group, optimizing the balance between comprehensive detection and resource efficiency.
3Measurement precision
If user identity information is fully analyzed to detect fraud rings, then detection accuracy improves, but user privacy is compromised
Solution Approach 1:
The patent introduces surrogate identifiers as an intermediary between raw user identity information and fraud analysis. These intermediaries preserve the grouping and pattern recognition capabilities needed for fraud detection while anonymizing individual identities. This resolves the contradiction by enabling accurate fraud detection through pattern analysis without exposing or storing sensitive personal information.
Data Source
AI summary
In some implementations, a computing device can be configured to perform impact-based fraud detection. For example, a computing device (e.g., a network server) can receive user submissions from user devices corresponding to traffic incident reports, point of interest (POI) ratings, product ratings, vendor ratings, and/or other crowd-sourced information. The computing device can identify groups of users based on the user submissions and various grouping criteria. The computing device can determine the impact of these user groups with respect to the targets (e.g., traffic in an area, ratings of a business, etc.) of their user submissions. The computing device can prioritize high impact user groups when attempting to detect fraudulent user group activity (e.g., fraud rings).