Research On Internet Consumer Finance Credit Risk Control Strategy Based On Xgboost

Under the influence of such ideas,Internet Plus,Inclusive Finance,etc,the traditional financial industry is rapidly integrating with current advanced Internet technologies,resulting in the financial industry being constantly subdivided.Internet consumer finance,which based on real consumption scenarios,has achieved rapid development in recent years and has broad market prospects.However,due to the widespread problems of moral hazard,adverse selection,short personal information,and lack of pre-lending multidimensional data collected in the Internet finance industry,there are various risks associated with the rapid development of Internet consumer finance.So how to rapidly and effectively control and prevent such risks,and how to establish a strong risk model and have a scientific risk control strategy are problems that need to be solved in the development of consumer credit business under internet background.First,this paper uses literature research method to draw lessons from the research results of domestic and foreign scholars,and points out the background,purpose and significance of this study,and summarizes the domestic and foreign research status of Internet consumer finance.Second,from three aspects of concept,characteristics,and models of Internet consumer finance,the relevant connotations of Internet consumer finance are explained,and four types of common risks of internet consumer finance are summarized.Then credit risk is selected as the risk subject of this study,Basing on traditional credit risk control,the demand and direction of new credit risk control based on machine learning under the background of big data are proposed.Thirdly,basing on the introduction of Xgboost's basic algorithm principles,the reasons for selecting the Xgboost model are given,and an internet consumer finance credit risk control strategy model based on Xgboost is constructed to provide theoretical support for empirical analysis of the following text.Finally,taking the test data of internet consumer finance company S as a research sample,and conducting an empirical analysis according to Xgboost based on consumer finance credit risk control strategy model that has been constructed in the previous section,By comparing with the training results of logistic regression model,SVM model and Gaussian Bayesian model,on the basis of comprehensive consideration of training duration,TPR evaluation indicators and interpretability of results,the superiority of the Xgboost model in screening key influencing factors that cause credit customers to have overdue performance is verified,and basing on the optimized credit risk control strategy,proposing corresponding strategies for credit risk control of S company.This paper proposes the consumer finance credit risk control strategy based on Xgboost model,it's systemic,timeliness,and risk forecasting capabilities have provided new ideas and suggestions for the design of risk control strategies for internet consumer financial companies.