Time-event data-based debt risk assessment method and system after credit card expiration, and application

The invention discloses a time-event data-based debt risk assessment method and system for an overdue credit card, and the system comprises two modules: a customer classification model and an overdue day number prediction model, and achieves the functions of depicting a user portrait of an overdue customer of the credit card, classifying whether the customer comes out of a prompt, and predicting the number of days of a good customer coming out of the prompt. A risk function learning model for time-event data is constructed, customers are classified according to customer risk conditions according to historical behavior data of the customers and whether the customers are overdue or not, and debt risks of the customers are comprehensively evaluated by predicting urging dates of the classified good customers. The problems of category imbalance, collection real-time analysis, comprehensive consideration of time and event indexes and the like existing in the overdue data of the credit card are solved, and compared