Abstract: Evaluation of borrower’s credit is an important step for the P2P lending company to control the risk, and it is of significance to company operation.This article analyzes the credit risk from the view of borrower. On the basis of reading a large number of P2P lending borrowers credit risk literature at home and abroad, we set up typical P2P lending platforms in the investigation, study and select the influence factors of P2P lending borrowers credit to establish the P2P lending borrower's credit evaluation system. After introducing the trainoss to optimize the BP neural network, we establish the BP neural network model according to the credit evaluation system. Then the model is trained and simulated by using the relevant data of the 104 borrowers in four representative P2P lending platforms. The training and simulation results show that the optimized model has better convergence and higher accuracy than the standard BP neural network model, and it can evaluate the borrower's credit more accurately. P2P lending platforms can use the model to filter the high quality of the borrower and control the borrower’s default risk, reducing the platform’s bad debt rate, so that the platform can make a stable operation. Therefore, the model plays a certain role in the risk control of P2P lending, so as it can promote the healthy development of P2P Lending industry.

Key words: P2P lending, BP artificial neural network, credit evaluation

摘要： 评估借款人信用是P2P网贷公司控制风险的重要步骤,对于网贷公司的正常运行有着极其重要的意义。论文参考商业银行信用指标体系并根据P2P网贷自身特点,建立了P2P网贷借款人的信用评估指标体系。根据建立的指标体系构建相应的BP神经网络模型,并利用一步正切法进行优化。然后选取具有代表性的P2P网贷平台的相关数据,对该模型进行训练和仿真,证明了该模型对P2P网贷平台的风险控制起到一定的作用。

关键词: P2P网贷, BP神经网络, 信用评估