The Construction Of The Risk Identification Model Of My Country's P2P Platform

In recent years,the rise of Internet finance has accelerated the flow of funds in the entire financial market.Various new economies have been rapidly derived.The domestic P2P industry has developed particularly rapidly.One of the participants in borrowing and lending.Due to its low threshold,P2P platforms make investors face great risks when investing.How to identify platforms with higher risks from various P2P platforms when investing is a major challenge faced by the government and investors.problem.This article aims to identify the credit risk of the P2P platform,so that investors can be more cautious in the investment decision-making process.This article first uses web crawler technology(such as octopus,etc.)to crawl out the basic information about the P2P platform,transaction information and Internet public opinion information,including 5 first-level indicators and 25 second-level indicators.Before data analysis,the SMOTE algorithm is used to balance the unbalanced data.Then use the supervised learning method in data mining technology,which includes SVM support vector machine,BP neural network and C5.0 decision tree.Through model comparison,the classification accuracy rate and the first type error rate from the test set(that is,high Risk misjudgment as low risk),the second type of error rate(that is,low risk is misjudged as high risk),C5.0 algorithm has the best classification performance in the risk identification process of P2P platform,and finally chooses C5.0 algorithm for features select.From the 25 variable indicators,the optimal 20 indicator systems for P2P platform risk identification were selected,and it was found that 20 indicators such as registration place,whether to join a supervisory association,average borrowing period,and bank depository are the most important factors affecting the risk of P2P platform Indicators,further use factor analysis to analyze the P2P platform risk evaluation system constructed by these 20 indicators,extract 6 common factors,and then calculate the P2P platform risk comprehensive score,and finally use the comprehensive score to rank 103 P2P platforms.Verify the rationality of C5.0 algorithm feature selection results and better risk prediction abilityof P2P platform.