Hyperight

Graph Machine Learning for Fraud Detection – Dan Saattrup Nielsen, University of Bristol/formerly Danish Business Authority

Bookmark

Premium content

Login or register to unlock the content

The Danish Business Authority started investing in graph databases and graph machine learning in 2018, and now has a machine learning lab of 15+ data scientists. I will talk about how we use graph analytics and machine learning to find tax fraudsters, money launderers and more.

Key Takeaways

  • Graph structure makes it easier to find networks of criminals
  • Centralised data access makes the graph approach more viable
  • Graph neural networks currently does not seem to outperform manually curated graph features

Add comment