Lending Behavior Of Bling

With the advance of electronic commerce and social network, P2P lending have a rapid of development in China. Compared with the traditional financial marketing, it has a lot of advantages, such as lower investment threshold, higher financing success rate and lower transaction costs. As those benefits almost beyond our imagination, P2P lending creating a new era of civilian investment. While the P2P lending is in high speed development, it also has brought many problems. Firstly, under condition of both legal and credit system are inadequate, asymmetry of information may easily cause moral risk and adverse selections. The P2P lending platform may suddenly bankruptcy, borrowers have to raise interest rate and investor irrationality is more serious. Secondly, many investors are concerned about higher interest rate in lending and the rising appetite for risk been highlight. Limited to useful tool, investor cannot identify the real signals from lending market. To some extent this also reflects the irrational behavior of investor.Initiated by background introduction and potential contribution, this paper systematically reviews the literature research about P2P lending. Furthermore, this paper makes a discussion about investor behavior in theory, partly revealed the cause and consequences of their irrational behavior. The three steps are as follow: Firstly, the transaction record on renrendai.com has certain regularity and it has advantages in the availability and integrity of data. Basing upon this data, this study discussed some factors that may affect investors'behavior on P2P lending market. In this step, it was found that the three type of learning algorithm by using random forest is best prediction effect. In the model of loan success rate, the prediction accuracy rate inside the sample and outside the sample is 97.82% and 97.839%. In the model of loan default rate, the prediction accuracy rate inside the sample and outside the sample is 96.43% and 95%.Secondly, we apply data mining methods to identity the priority of various factors. We found hard information of lenders perform well in identify sources for risks. But it is also found that some of important factors affecting the default has not been considered by investors. Such as working time, lending rate, lending term, amount of loan are both of the important factors in both of two model. We could say that this result represent the part of the investor's rationality. In the other hand, the important factors which may affect the default rate, including the work certificate and proof of income. This simply means the irrational behavior of lending marking in China.In addition, there has many of variables in our data, including the dummy variables as high as 36. In order to ensure the accuracy and interpretability of the model, we use the method of variable selection. The paper incorporates Lasso, a high dimensional index selection method with logistic regression to establish another success rate and financial risk model. We found that the limitation of time and energy, the investor cannot seek the most appropriate intersection of risk and return. Blinding investor is fairly common problems on P2P lending market.