The Research And Application Of Financial Risk Model Before The Loan Based On BP ANN

In the complex environment, every enterprise has risk. Commercial bank is the special enterprise that runs the financial assets.The most important risk of the commercial bank is the loan risk. The essence of the loan risk is the uncertainty of safety factor. In the process of the loan risk control, the point is rising the veracity of the assess of Probability of Default, so it is very important to get a right grade of the enterprise financial risk before loan. It can help the manager reject the enterprises which have a high financial risk, thus can help the bank reduce the lost.The management of loan risk starts late in China. We still use the old methods to assess the loan risk, so we sometimes can not get a exact result and the information is always incompletely. The neural network can deal with the information fast and can learn well by itself, it can increase the learning capacity with the change of the environment, so it begins to be used in the access of loan risk.This paper chooses the BP neural network, which can increase the learning and storage capacity, to complete the training and learning process of the model. The model contains the processes of allocation, training and testing. All the data used in the model is collected in the loan data of Shenyang Branch SPDB, which contains 13 indicators (Asset-liability ratio, property ratio, the multiple of interest safeguard, the current ratio, fast acting ratio, net assets yield, return on capital and profit rate of main business, net profit margins, the account receivable turnover, inventory turnover, total asset turnover and income growth).This paper uses the matlab software to train and test the model, and we get a perfect output, it can declare the model can be used in risk control before loan.