Fraudulent Network Detection via Graph Analysis
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current methods fail to effectively detect fraudulent publisher networks in Internet advertising, as cybercriminals employ sophisticated techniques to simulate traffic across multiple websites, evading detection by maintaining low revenue thresholds on individual sites while collectively generating significant revenue.
Innovation Solution
A system and method utilizing a graph generation system, network identification system, and network scoring system to analyze data from multiple entities, generate representations of similarities, identify candidate fraud networks, and determine fraud scores to detect and flag potentially fraudulent networks.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If traditional detection methods are used to identify fraudulent websites, then individual fraudulent sites can be detected, but cybercriminals evade detection by employing multiple websites that individually generate low revenue but collectively yield significant money
Solution Approach 1:
The patent combines multiple previously independent detection signals (traffic patterns, revenue data, website characteristics) into a unified network-level analysis framework. By merging these disparate data sources and analyzing them collectively rather than individually, the system can identify fraudulent networks that would be invisible when examining single websites in isolation.
Solution Approach 2:
The patent creates a multi-functional detection system that simultaneously performs multiple functions: collecting data from diverse sources, analyzing network relationships, identifying fraudulent patterns, and generating actionable intelligence. This universal approach replaces multiple specialized detection tools with a single integrated platform that handles the complexity of modern fraud networks.
2Reliability
If analysis is performed on individual websites, then detection thresholds can be set based on single-site metrics, but this allows fraud networks to operate undetected by keeping each site below detection thresholds while collectively generating significant revenue
Solution Approach 1:
The patent transitions from analyzing fraud in a single dimension (individual website metrics) to multiple dimensions by incorporating network-level relationships, cross-site traffic patterns, and aggregated revenue data. This dimensional expansion allows the system to detect fraud that operates below thresholds in any single dimension but becomes apparent when viewed across multiple dimensions simultaneously.
Solution Approach 2:
The system performs preliminary data collection and network mapping before fraud fully manifests. By continuously gathering and analyzing data from multiple sources in advance, the system builds a baseline understanding of normal network behavior, enabling it to detect deviations and fraudulent patterns before they generate significant revenue loss.
Data Source
AI summary
The present teaching generally relates to detecting fraudulent networks. First data associated with a plurality of entities may be obtained, and a representation characterizing similarities among the plurality may be generated. Based on the representation, at least one entity cluster may be identified as corresponding to a candidate fraud network. A score associated with each of the at least one entity cluster may be determined, where the score indicates a likelihood that a corresponding entity cluster represents a fraud network, and at least some of the at least one entity cluster may be identified as a fraud network based on the score.