Credit card bogus transaction dichotomy model construction method and bogus transaction judgment system
Patent Information
- Application Number
- CN202310586121.8
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2023-05-23
- Publication Date
- 2024-11-26
AI Technical Summary
The existing technical methods for monitoring false transactions of commercial banks' credit card are insufficient in accuracy, coverage and timeliness. Rules and decision-making based on expert experience rely on human experience and are difficult to unify. Model prediction based on supervised learning algorithms is not accurate in discrimination of real transaction labels. , and limited to historical data, it is difficult to fully cover false transaction types.
The construction method of the binary classification model of false credit card transactions is adopted, and the credit card transaction details and customer data during the billing cycle is collected, and the semi-supervised learning algorithm and XGBoost algorithm are used to build the binary classification model of false credit card transactions is used, and expert review and group stability indicators are used. PSI verifies model performance and improves the discrimination ability of false transactions.
It improves the accuracy, coverage and timeliness of false transaction monitoring, and enhances the ability to judge false transactions on credit cards.
Smart Images

Figure CN119027226A_ABST
