Research On P2P Network Lending Risk Early Warning Model

Since 2007,the first P2P network lending platform pat loan in China was established in Shanghai,then kinds of P2P lending platforms springing up.Because of the low operating threshold and high flexibility characteristics,P2P network attracts more and more people to join in.Though the development of P2P lending industry in China is getting faster and faster,the influence of momentum is also became huger and huger because of the lack of supervision.Risk control in P2P lending market is so immature,which caused a huge hidden trouble for the entire industry.At this moment,the fragile ability to control the risk became a common fault in Chinese net loan platform.Therefore,in order to make the development of whole industry become rapid and steady,the P2P network lending industry must improve its ability in risk control by building an effective risk control system.This thesis takes the P2P network lending risk warning model as the research subject,the main research content is as following: Analyzes the related concepts of P2P network lending;determine the net credit platform quantitative indicator of risk factors by combing commercial credit index system with the survey of network risk;based on the principles of index screening and cluster analysis methods to eliminate indicators,selected 6 indicators as the main indicators of risk control,and then use the principal component analysis to extract risk factor,risk and calculate the score.Use BP neural network construct the model,of which the six selected quantitative index as input vector,the risk comprehensive value as the output vector;Choose the data of 30 platforms in 15 months,including 20 platform data as a training set trains the neural networks,the remaining 10 data set as test platform and analysis the result of the experiment;propose the risk control of the P2P lending industry.