P2P Platform Risk Identification Scheme Planning Based On Soft Information

With the rise of Internet finance industry,P2 P has developed rapidly in China.P2 P network loan is a personal to personal financial model,which not only promotes the development of the financial industry,but also better arranges the idle funds in the society.However,because the relevant domestic laws and policies are not perfect,the supervision is not strong,and the risk early-warning mechanism is not mature,in recent years,P2 P platforms frequently "explode thunder","run on the road",close business or withdraw cash difficult phenomenon appears more and more frequently,the platform can not realize the phenomenon of cashing is common,making the interests of investors suffer losses.Therefore,it is necessary to identify and prevent the risks of P2 P network loan platform.However,in order to quickly obtain huge profits,some online loan platforms have a large number of transaction data omissions and fraud,which greatly reduces the risk identification function of P2 P platform by hard information.It is not enough to rely on the basic information of online loan platform and the hard information such as transaction information for risk identification.Therefore,to identify the risk of online loan platform,we need to give full play to the soft information such as online public opinion and senior management information Effect.Based on this,this paper takes P2 P network loan platform as the research object,carries on the risk identification to P2 P platform,and according to the result of the scheme planning,puts forward the suggestions for the government,P2 P network loan platform and investors.First of all,the related literature of Internet financial risk at home and abroad is combed,and the relevant theories of risk assessment and early warning model research are analyzed;secondly,the theory of information asymmetry and the problem of information asymmetry in the Internet loan industry are introduced,and the commonly used machine learning algorithms are introduced;then,combined with the risks faced in theoperation of P2 P Internet loan platform in China,according to the risk impact Response factors select risk identification variables,and add soft information variables of online loan platform to the identification variables,including online public opinion information and senior management information of the platform.After adding soft information variables through logical regression model and support vector machine model verification,the accuracy of risk identification is improved,and the accuracy of risk identification of support vector machine model is higher;finally,further proposed Combined model lightgbm-lr,and verify the combined model for online loan platform risk identification,not only the accuracy is higher than support vector machine,but also the results of better interpretability.According to the research results of this paper,the registered capital,number of senior managers,reference income,business hours,online public opinion,senior management information and other factors of the online loan platform will have an impact on the risk of the platform,and under the condition that other factors remain unchanged,the lower the platform evaluation and experience,the fewer the number of senior managers,the lower the education background and the shorter the working time,the more likely the online loan platform is to run Road and other risk issues.According to the conclusion of this paper,the platform puts forward some suggestions for government supervision,investors' investment and the standardized operation of online lending platform.