Credit Loan Default Prediction Based On Data Mining

Abstract

With the continuous development of economy and the improvement of people's level, personal credit loan develops rapidly. Because the problem of credit loan default is becoming more and more serious, the development of credit loan business needs an accurate prediction system. Banks have a lot of historical data, using data mining technology, from the basic information, social relations, consumption behavior, such as address information as much as possible in the massive amounts of customer data mining on the information of the borrowers, summed up the main factors influencing the personal credit risk prediction, and based on this model can effectively predict the personal credit risk related. The results show that the prediction accuracy of XGBoost model is higher than that of Logistic Regression model and Random Forest model.