Research on Credit Risk Early-Warning for Listed Companies in Chengyu Economic Zone Based on Best Fuzzy Support Vector Machine

摘要

Taking listed companies in Chengyu Economic Zone as an example, this paper introduces the fuzzy algorithm into support vector machine (SVM), constructing the model of fuzzy support vector machine (FSVM) for Credit risk early-warning, which based on four different kernel functions (linear, polynomial, sigmoid and Gauss radial basis) are compared as well as compared with traditional statistical models and other artificial intelligent models. The result of investigation illustrates that FSVM based on Gauss radial basis kernel function is not only superior to that based on other three kernel functions, but also better significantly than traditional statistical models and other artificial intelligent models.