信用评级方法模型的研究综述与展望

The Research Review on the Credit Rating Methods and Models and the Outlook

本文对信用评级的几大类方法如信用评级的定性分析如层次分析法、统计分析方法如Probit模型、神经网络与支持向量机方法如支持向量机(SVM)、其他方法如投影寻踪法等相关研究及应用成果分别进行了阐述。本文认为,各模型具有内在的优点和缺陷,应用多种分析方法相结合对企业进行信用评级,有助于提高模型的预测价值和结论的可解读性。

The paper respectively expounds the related research and application results of the major kinds of credit rating methods such as the qualitative analysis of credit rating like Analytic Hierarchy Process （AHP）, statistical analysis methods like Probit model, neural network and support vector machine methods like Support Vector Machine （SVM）, and other methods like projection pursuit method. The paper believes that each model has its own inherent advantages and defects, so a variety of analysis methods should be combined to make the credit rating for the enterprises, which helps to improve the predictive value of the models and the rationality of the conclusion.