Icon

Isolation_​Forest_​for_​Fraud_​Detection_​Deployment

Fraud Detection: Model Deployment

This workflow, the deployment workflow, reads the trained model, as well as the new transaction and applies the model to classify it. Like in the training workflow we use a Rule Engine node to apply the custom threshold. In case a transaction is classified as fraudulent the workflow sends an email to the fraud department.

Fraud Detection: Model DeploymentThis workflow, the deployment workflow, reads the trained model, as well as the new transaction and applies the model to classify it. Like in the training workflow we use a RuleEngine node to apply the custom threshold. In case a transaction is classified as fraudulent the workflow sends an email to the fraud department. Model Deployment Create H2O FrameRead onetransactionApply IsolationForest to get Mean Length Send Email in case of fraudulant transactionCreate KNIMETableClassify Based onMean LengthRead Isolation Forest Table to H2O H2O Local Context Table Reader H2O IsolationForest Predictor Table Rowto Variable CASE SwitchVariable (Start) Send Email H2O to Table Rule Engine Model Reader Fraud Detection: Model DeploymentThis workflow, the deployment workflow, reads the trained model, as well as the new transaction and applies the model to classify it. Like in the training workflow we use a RuleEngine node to apply the custom threshold. In case a transaction is classified as fraudulent the workflow sends an email to the fraud department. Model Deployment Create H2O FrameRead onetransactionApply IsolationForest to get Mean Length Send Email in case of fraudulant transactionCreate KNIMETableClassify Based onMean LengthRead Isolation Forest Table to H2O H2O Local Context Table Reader H2O IsolationForest Predictor Table Rowto Variable CASE SwitchVariable (Start) Send Email H2O to Table Rule Engine Model Reader

Nodes

Extensions

Links