Machine Learning for Fraud Detection with Graph

Machine learning has helped detect fraud, whether it is payment card fraud, application fraud, or incentive fraud. Often times, fraud can be detected by looking at the topology of a graph. For example, how many transaction/information paths connect a user to a blacklisted account? Features like these are more easily calculated using graph databases than traditional relational/nonrelational databases. In this session, we look at example queries that can extract features from a graph topology for machine learning.

• Xinyu Chang, Director Customer Solutions, TigerGraph
• Nalu Zou,  Solutions Architect, TigerGraph 

Download the session slides here.