Optimization Of P2P Personal Credit Risk Assessment Model Based On Data Mining Technology

With the rapid development of P2 P lending business in China,many P2 P platforms have experienced bad debt rate and even platform running.These phenomena are caused by the fact that the P2 P platform does not effectively evaluate the individual credit risk of investors.This paper explores the optimization of P2 P personal credit risk model through data mining technology.In this paper,through the 11017 loan data of the “patting loan” platform,9206 loan data and 110 loan variables are obtained after data preprocessing and variable analysis,and the logistic regression model,SVM class model and decision tree class model are respectively modeled.training.After the model prediction,the model training efficiency and the model prediction effect indicators are comprehensively evaluated,and a single data mining model with good model evaluation(XGBoost model and logistic regression model)is selected.In the second stage,the XGBoost model and the logistic regression model are merged into the Stacking model to obtain the final model optimization form,and the model training and model prediction are carried out.The data mining technology is optimized for the P2 P personal credit risk assessment model.Through the research in this paper,we also get the research conclusions and research implications for P2 P platforms,investors and regulators.