Personal Credit Risk Assessment Of LC Company

With the development of Internet technology and economic,society needs financial services.In recent years,China's internet lending platform has experienced from explosive growth to the steady development stage.The power of development of internet lending platform is that financial service system can not satisfy the demand of enterprises and individuals,but the requirement is remedied by the internet lending platform that is convenience and quickly.The core issue of network lending platform is risk control of financial services when lending platform provided financial services for customer.There are many methods for credit risk assessment,but the most wide used is that the estimate model based on personal assets,personal credit data and other data.This paper introduces a method of constructing network platform for personal loans risk assessment model.This article's focus is that individual credit evaluation of network lending platform,and the study prediction model,and the processed model's method.The paper's study sample data is the Lending Club data that in 2016 fourth quarter.The research is that using statistical learning method to establish model.The first,to clean and convert data and extract key feature.Then,the risk evaluation model is made up by Logistic regression,xgboost model and combined model.The method of check predict result's correction is that the model is applied in the test data to check K-S test and AUC value.The final conclusion about the personal credit evaluation internet lending platform is that using the Logistic regression method can be explained and xgboost model's predict result is more exacter.The combined model have the characteristic that it is higher of predict result's correct rate and explained than Logistic regression method and xgboost model.