Study On Development Of P2P Lending Based On Data Mining

Financing is an important problem for SMEs and the general public for a long time.P2P lending reliance on the development of the Internet provides a quick solution to this problem.With the promotion of the Internet economy,the number of platforms with various backgrounds and sizes has increased.And platforms have different strengths and qualities.Rad debts,suspensions and running away are all possible problems.Therefore,research on the factors that affect the development of the P2P online loan market is a subject worthy of attention.First,this paper integrates the concerns of investors in P2P online lending,the potential risks of the platform,and macroeconomic factors.And then,relevant data sets were collected.Next,some data mining methods were used.In the end,empirical analysis was conducted.It mainly includes the following three points:Random forest,bagging and survival analysis methods in data mining are used in order to explore the impact of "survival" or "death" on the P2P online loan platform and provide a theoretical reference for investors and regulators aiming at the characteristics of "life is short" appearing on the P2P network platform,Rate of platform is one of the most concerned.Therefore,actual data set as research sample,classic financial time series ARIMA model and the particle swarm optimization algorithm were combined to realize the prediction of the short-term profitability of a representative P2P online loan platform.P2P online loan will be affected logically by macroeconomic changes.Some data set related to P2P online loan market and macroeconomic were collected.And then,Granger causality test and cointegration model in data mining were adopted in order to explore impact between the P2P online loan market and the macroeconomic.In general,this paper has a certain theoretical significance and practical significance.Application fields of data mining technology were extended.A reference was provided for investors and regulators making decisions.And some data sets and experimental results also provide information for follow-up scholars' research.