Research On The Combination Prediction Of P2P Lending Default Based On Combination Weighting Function

The development of the Internet has given birth to a new model of P2 P lending.In this mode,the borrower's required funds are borrowed through the network platform,which lowers the threshold for borrower to borrow funds and improves the efficiency of using funds,but it is easy for the borrower to default.Therefore,correctly predicting the defaults of P2 P borrowers is of great significance to the healthy development of P2 P platforms.This paper analyzes in detail the research results of domestic and foreign scholars on the default prediction of P2 P lending borrowers,and discusses the influencing factors and prediction methods of the default of P2 P lending borrowers.From the perspective of the environment of P2 P lending market and the statistics of P2 P lending,this paper makes an in-depth analysis of the causes of the defaults of P2 P borrowers and summarizes the types.Considering the changes in the borrower's mind,this factor was added to the borrower's default influencing factors,and a theoretical model of the borrower's default influencing factors was constructed.Therefore,based on the data mining theory,this paper studies the application of SVM,decision tree and logistic regression in the prediction of P2 P lending default.In order to get better prediction results,this paper proposes a P2 P lending default combination prediction model based on a combination weighting function.The research focuses on the following issues:First,the paper uses a combination of qualitative and quantitative methods to analyze the influencing factors of P2 P lending borrowers' default.Firstly,it analyzes the influencing factors of borrower default from the theoretical level,and then analyzes the influencing factors of borrower default from the data level.From a qualitative and quantitative perspective,the results obtained are more scientific,making up for the singularity of using only quantitative methods,and being more intensive.Second,construction of the theoretical model of influencing factors of borrower default in P2 P lending.Based on a detailed analysis of the influencing factors of borrowers' defaults by domestic and foreign scholars,the influencing factors of borrowers' defaults in P2 P lending behaviors are divided into individual factors,interpersonal factors and characteristics of loans themselves.Based on this,the inner emotional changes of the borrower are added.Through the analysis of SOR theory,Heider balance theory and Knowledge Attitude/Belief Practice theory,the theoretical model of borrower default behavior change in P2 P lending behavior is constructed from multiple perspectives.This paper explores the influencing factors of borrowers' default from the theoretical level and provides a new research idea for P2 P lending.Third,construction of a P2 P lending default prediction model based on the combination of weighting function.On the basis of the data mining theory,the paper establishes SVM,the decision tree and logistic regression of three basic models,then proposes combination weighting function method,which focuses on deviation of real value and combination forecast,and the deviation between predicted values and the combination forecast model.and pay more attention to the weight distribution,has a certain theoretical and practical.In the experimental research of P2 P lending,few scholars pay attention to the problem of dealing with data imbalance.This paper uses SMOTE method to unbalance the P2 P lending data.The experimental results show that the effect of the model after the unbalanced treatment is significantly better than that before the treatment,and the model can predict the defaulting borrowers more accurately.The final experimental results show that the prediction effect of the combined model is more ideal than that of the single model.