Skip to content

Fraud Detection

A Taipy Application that analyzes credit card transactions to detect fraud.

Try it live Get it on GitHub

Understanding the Application

This application shows a list of credit card transactions. The user can select a date range to predict fraud. The application will then use an XGB model to mark potentially fraudulent transactions in red or yellow.

List of Transactions Page

The user can select a transaction to see an explanation of the model's prediction, as well as the client's other transactions.

Prediction Explanation Page

The user can also choose the threshold of the model. The threshold is the model output above which a transaction is considered fraudulent. The user can select the model according to the displayed confusion matrix and by looking at False Positive and False Negative transactions.

Threshold Selection Page