02_​Fraud_​Detection_​RF_​Model_​Deployment

0 ×

We read the trained model, as well as the new transaction and applies the model to classify it. We use a Rule Engine node to apply a threshold. In case a transaction is classified as fraudulent the workflow sends an email to notify of a fraud.

This workflow demonstrates how we can use the trained Random Forest Model on new data by performing the following steps:
1. Read the model and new data
2. Apply the model on the new transaction

Nodes

• CASE Switch Start1 ×
• Component Input1 ×
• Component Output1 ×
• Model Reader1 ×
• Random Forest Predictor1 ×
• Show all 10 nodes

Extensions

• FeatureKNIME Base nodes
• FeatureKNIME Ensemble Learning Wrappers
• FeatureKNIME Javasnippet

Links

• No links available