Isolation Forest for Fraud Detection: Model Deployment

This workflow reads the trained isolation forest model, as well as the new transaction and applies the model. Based on the isolation number (mean length) and a threshold the workflow detects fraudulent transactions and sends an email, if the isolation number is higher than the threshold.

This is a companion discussion topic for the original entry at