Fraud_Detection_DBSCAN_Deployment
Fraud Detection: DBSCAN Method Deployment
We read in all our saved models and tables to prepare for the similarity search. Read the new transaction and add it to the normalize (apply) with the saved model. Perform the similarity search and classify whether or not the transaction is potentially fraudulent or not. If it is, an email can be sent to notify.
This workflow demonstrates how we can apply our models on new data by performing the following steps:
1. Read the models, dbscan table, and new data
2. Apply Normalizer and Perform a Similarity Search on Incoming Transaction
3. Mark Outliers
Nodes
Extensions
Links
- No links available