The Credit Risk Prediction of the Small and Medium-Sized Enterprises Based on GA-v-SVR

Making the science assessment and prediction of the credit risk of the small and medium-sized enterprises (SME) is a significant part of risk management of commercial bank. This paper firstly integrates Genetic Algorithm (GA) with v-SVR model, creates the credit prediction model, GA-v-SVR, and then builds the SME credit risk indicator system. Using principal component analysis method screens out the major factors of the credit risk, it then considers those factors as the input indexes, and chooses credit default as the output indicator. The model is employed to train the first group sample, and the rules of credit risk recognition are obtained. Finally, we utilize these rules to predict the credit of the second group sample, and acquire a rather accurate result so that can prove the feasibility and validity of the model established by this paper.