Research on the Forecasting of Enterprise Credit Scoring Based on SVR Model

Abstract

In the competitive market environment, enterprises have to face all kinds of risks at any time. Among them, credit risk is one of the most challenging risks. As the main analysis method of credit risk assessment, how to establish and improve the credit scoring model of the company is of great significance to the enterprise. This paper proposes a support vector regression (SVR) model with kernel function to solve the forecasting problem of corporate credit scoring. In this paper, the model is applied to a group of basic information data sets of enterprises with specific credit score values, and we study the prediction effect of the model . In the experiment, linear regression model and neural network model are selected to compare with SVR model. The experimental results show that the accuracy and efficiency of SVR model are better than the other two models. Compared with the traditional expert scoring method, the SVR model can automatically predict the company's credit score according to the existing information data of the enterprise, which not only saves cost but also improves efficiency. Therefore, the model has a high reference value for dealing with the forecast of enterprise credit rating.