The Application Of Statistical Methods In Internet Financial Default Risk Identification

The rapid development of the Internet has changed the way we live,especially in recent years,the Internet and financial has combined.Relying on the Internet,Big data,Cloud computing and other technologies Internet financial has deeply affected the traditional financial industry and changed people's way of life.But based on the characteristics of the Internet and the system is not perfect,Internet finance also has a certain risk.Strengthen supervision and reduce risk has become the most big challenge of the Internet financial industry.This paper studies the credit risk of Internet finance,and aims to find out how to predict the existence of default risk by analyzing the data of Internet Finance Company,and reduce the loss of the company to the greatest extent.In this paper,through the study of relevant achievements at home and abroad,combined with the development of China's Internet finance,the use of statistical knowledge and machine learning methods for modeling analysis.A total of 1000 samples and 29 variables of financial data were selected as the object of study.Before the model is established,the original data is processed and descriptively analyzed.The data are divided into two parts: training set and test set.In order to find the most suitable method,based on the comprehensive analysis of the advantages and disadvantages of the existing model,this paper establishes the Decision tree model,the Naive Bayesian model and the Logistic regression model respectively,and uses the test set data to verify the confusion matrix.In order to improve the performance of the model the adaptive boosting was added to the Decision Tree model and the Laplace correction was added to the naive Bayesian model,and then the F-index and ROC curves was used to find the best model is Logistic regression.Finally,we summarize the results of the study and analyze the problems we found and give recommendations.