Consumption staging default probability model based on survival analysis

Abstract

The invention discloses a consumption staging default probability model based on survival analysis. The model is characterized in that the number of consumption staging repayment periods is regarded as discrete survival time, the discrete survival time is added into the user attribute characteristics and then the discrete survival time is fused into a default model, the relationship between a sample survival result and the survival time and the user attribute characteristics is researched, and a model capable of predicting the default probability of the consumption staging loan user in any future period is established by using an xgboost algorithm; a danger function of survival analysis is obtained through the default probability model, the default probability of any loan user in any period is predicted based on the danger function, and the future default risk of the loaned user is evaluated at the same time. According to the invention, the possibility that the user is overdue betweenthe end of the observation period and the full period neglected by the traditional model is covered, so that the future risk is estimated more accurately, and the financial institution can develop theconsumption staging business more healthily and continuously.

Description

technical field [0001] The invention is based on the data characteristics and repayment performance of consumption installment loan users, regards the repayment period of consumption installment as discrete survival time, applies survival analysis related methods to the default model, and uses the Xgboost algorithm to obtain A model for predicting the default probability of consumer installment loan users in any period in the future. Background technique [0002] In recent years, the proportion of total retail sales of social consumer goods in GDP has shown an upward trend, and consumption has become more and more prominent in driving the economy; the proportion of consumer credit, such as loans from financial institutions, housing consumption loans, and car loans, has also continued to increase, playing an increasingly important role in supporting the economy. Obviously, consumer credit is not limited to major banks. Major e-commerce companies and licensed consumer finance ...

AI Technical Summary

Problems solved by technology

In addition, the traditional consumption installment default model does not treat the difference in the overdue risk of different users in different installments, but in fact, the early overdue probability of high-risk users is higher than the late overdue probability

Images