The Comparative Research Of Personal Credit Assessment Model Based On BP Neural Network And SVM

Credit risk assessment is the core content of banking risk management. Thestudy of global banking crisis by the world bank shows that, the most commonreason of causing bank failure is credit risk. Personal credit assessment is the akey process of personal financial business and credit approval in commercialbank’. Its the core of personal credit risk management. It’s urgent to get scientificquantiifcation of personal credit and set up perfect and auto personal creditevaluation system to expand individual credit business and improve themanagement efficiency. We can accurately know the status of the personal credit,correctly identify the risk of personal credit, and promote the optimizedallocation of credit resources through the science assessment of personal credit.BP Neural network is a self-organizing, adaptive and self-learningnonparametric method, It is not strict with the distribution of sample data，notonly has nonlinear mapping ability, but the high prediction accuracy. Theseadvantages make BP neural network achieve rapid development in the field ofcredit risk assessment. Support vector machine (SVM) is a kind of machinelearning algorithms. It can obtain a minimum error classiifer through training thelimited samples. It has a good classification and generalization ability to achievethe correct classification of the test. SVM can well solve the small sample,nonlinear, high dimension of complex problems.The paper established the index system of personal credit and set uppersonal credit evaluation model based on BP neural network and SVM by using’public data sets. Its found that the effect of BP neural network and SVM is goodin personal credit evaluation through the comparison, and the SVM has higherrobustness. The SVM show better predict precision when the sample sizes havebig discrepancy.