Using Graph Databases & Analysis for Collusion Fraud Detection

11:10 am - 11:30 am

Objective: Guidance
Audience Level: Beginner/Intermediate
Session Type: Presentation

We propose an approach using graph databases and analysis to detect complex relationships of fraud users. Graph database systems map the shortest path between nodes and define cliques. Sets of nodes in a clique become closed ecosystems or have fewer edges than population average. Applying this method improves our ability to detect and stop fraudulent behavior, versus using relationship databases.


, Risk Strategy Analyst, Uber Inc