Research On P2P Credit Personal Credit Evaluation Model Based On BP Neural Network

In an era of rapid development of the market economy, personal credit as a necessary prerequisite for personal consumption, expanding domestic demand and stimulating financial credit sector plays a vital role in the development. Not only that, as a carrier of personal behavior, personal credit, has gradually been recognized, is a kind of personal morals, necessary means of maintaining economic and social order. The importance of personal credit scoring can be seen. However, specialization of individual credit rating system has been entered in the West even at the stage of industrialization of China for P2 P personal credit evaluation system of construction has just started.P2P lending was introduced to China in 2006, and then P2 P lending companies in China to obtain the explosive development of number by 2008 in dozens of homes quickly grew to more than 300 today and fast webified. But the lag due to regulatory and legal regulations will bring substantial economic risk to lenders and, therefore, individual credit assessments of P2 P lending model are extremely important. P2 P credit evaluation model for personal credit helps lenders measure of the debtor’s credit quality, the results can be used as an important indicator of whether lending is one of.2011 China’s banks Regulatory Commission released to the Banking P2 P lending credit risk warning, saying P2 P lending service agency there are a large number of potential risk, requiring banking institutions to take effective measures, risk monitoring and prevention. Therefore, in the practice of P2 P lending, on the borrower’s credit is very important to make a real and objective assessment of practical significance.The paper base on BP neural networks and establish a P2 P loan personal credit evaluation model. Mainly to do the following work:Analyze and comb the concepts and principles of BP neural network, and further summarized the ideas, principles and method of BP neural network, explains the basic method of BP neural network. Based on this in-depth analysis of the advantages and the problems faced by BP neural network.Summarizes the features of the history of credit, personal credit and personal credit rating. Analysis of the features of P2 P lending and borrowing from a credit point of P2 P lending is described.The establishment of personal credit evaluation index system of P2 P lending. Crossover frequency analysis is applied, some valuable statistics for each credit index description and qualitative analysis, based on individual credit rating system provides relatively intuitive basis for interpretable. And according to the available experimental data, establishing personal credit evaluation index system of P2 P lending.Summarized the existing research methods and research ideas to solve personal credit rating, based on a different method of data processing, are given personal credit rating classification, and for each type of approach were elaborated, and summarizes the advantages and disadvantages.In the tradition of BP neural network based on adjusting the angle of the right value, introduce additional momentum and elastic gradient descent method, proposed a new combination of two methods of weight adjustment method, and proposed a new weighting method employed in BP neural network.To P2 P lending(http://www.renrendai.com) Web site collects data for experimental data sets, this new weight adjustment method and the traditional method of BP neural network weights adjusted, additional momentum and elastic gradient descent method for multiple sets of experiments. New algorithm accuracy and convergence rate than traditional weight adjustment algorithm of BP neural networks. Innovation of this paper has the following aspects:Analysis of the features of P2 P lending, summarize existing data on P2 P lending site indices ’ influences on personal credit, summarized from the lender’s perspective a P2 P loan borrowing process.Setting up a P2 P loan personal credit evaluation indexes, each index by the crossover frequency methods effect on personal credit.From weight value adjustment, introducing additional momentum and elastic gradient descent method proposed one for P2 P personal credit evaluation model of combining additional momentum and elastic gradient descent method of weight adjustment.