Inquiry Response: Maturing Beyond Outlier-Based Fraud Detection

Inquiry:

We’re exploring ways to mature our fraud detection models. At the moment we rely on outlier-based detection, but we don’t have a ton of label data. How can we advance our outlier testing until we generate enough labels to build ML models? What else might you suggest?

Response:

THE BIG IDEAS:

  • Local outlier factor and isolation forest
  • Expand the data
  • Other ways of thinking about the problem

LOCAL OUTLIER FACTOR AND ISOLATION FOREST

You have traditional outlier-based approaches

Become a Research & Advisory Network Client to see the rest. Contact us to get started

Log in: