| Online P2 P lending platform,with its characteristics of low threshold,limited conditions and high yield,quickly occupied the Internet financial market.With the rapid development,the risks are gradually highlighted and the platform problems are gradually exposed.The frequent occurrence of platform running,liquidation,explosion and other problems not only caused huge losses to investors,but also posed difficulties to regulators.The survival of the online lending platform is influenced by many factors,the average lending rate of return related to its survival and development is also determined by many factors.It is of great significance to analyze the survival of online P2 P lending platform and dig out the key factors that affect the survival and development of the platform.Bayesian learning is a statistical method that sets a priori distribution of parameters and obtains a posteriori distribution based on sample information to obtain the overall distribution.Bayesian learning theory uses prior distribution probability to express all forms of uncertainty,and uses probability rules to realize learning and reasoning process.Bayesian learning method by using prior distribution and sample information to get the posterior probability of random variables,which improves the quality of statistical inference,even in the case of less sample data,more reliable analysis results can be obtained according to sample information and prior experience,to realize the robustness,expansibility and flexibility of the model.In this research,by collecting the statistical indicators and data of P2 P online lending platform released by www.wdzj.com,www.p2 peye.com,We combine Bayesian method with traditional Cox model and quantile regression to analyze the survival of online P2 P lending platform under the Bayesian framework and study the influencing factors;Construct Bayesian Network structure to classify and predict online P2 P lending platform;Use the entropy weight method and grey relational degree theory to evaluate the credit platform comprehensively.The main contents are as follows:(1)In view of the short life cycle of online P2 P lending platform,such as runaway,liquidation and other problems,the survival analysis of online P2 P lending platform is carried out.Firstly,variables are selected by Lasso-Logistic regression method.Then,the parameters of Bayesian Cox model are estimated by MCMC simulation,and the results are compared with the classical survival analysis Cox model.(2)we study the average lending rate of return of the online P2 P lending platform,and study the influencing factors of the average lending rate of return.The Internet financial data such as online P2 P lending platform has the characteristics of peak tail,and Asymmetric Laplace Distribution(ALD)is strong daptability to this kind of data.Therefore,we use Bayesian quantile regression with error obeying ALD was used to analyze the impact of each index variable on the average lending rate of return of online P2 P lending platform under different quantities.(3)In the mixed online P2 P lending industry,how to distinguish and choose the platform with good living conditions among many types of online lending platforms,and make scientific investment and lending to meet the needs of investors and borrowers is a topic worthy of attention in the research of online P2 P lending.In this paper,we study the Bayesian Network structure of the online P2 P lending platform,reflect the relationship between indicators through the network topology structure,and classify and predict the platform to better judge the survival status of the online P2 P lending platform,so as to achieve the effect of survival early warning.(4)In view of the survival problems faced by online P2 P lending platforms,we make a comprehensive evaluation of development of some online lending platforms.According to the objective evaluation of the data information,the Entropy Weight method,combined with the Grey Relational Degree to analyze the relationship between indicators,we calculate the comprehensive score of the platform development,carry on the comprehensive evaluation and ranking of the platform,to provide a reference value for measuring the survival status of the platform.In this paper,we use three models under the Bayesian framework to analyze the survival of online P2 P lending platform,study the factors affecting the average lending rate of return and classify the survival situation.Empirical research shows that Bayesian method can more intuitively explain the fitting effect of the model for P2 P data,which is easy to understand;and it can effectively use the data information to analyze the survival and development of P2 P platform.Finally,the comprehensive evaluation and ranking of online lending platforms are carried out.We want to provide a reference for investors,borrowers,regulators and the platform itself to provide survival risk identification and classification early warning method. |
PDF Full Text Request