摘要
债券是企业的重要组成部分,对债券进行信用评级能够帮助企业及时预防风险、帮助投资者正确评估企业的相对价值。随着中国公司债的发展,传统的评级方法已经不再适用。近年来,决策树、神经网络、支持向量机等机器学习算法已经走进信用评级领域,取得了较好的成效。本文以2016年上市公司发行的公司债为例,利用支持向量机算法对债券进行信用评级。并通过支持向量机中不同核函数的比较,选择最合适的核函数,发现在债券信用评级的问题中利用线性核函数进行分类的效果更好。
Bond is an important part of the enterprise,the credit rating of the bonds can help enterprises to prevent risks in a timely manner,to help investors to correctly assess the relative value of the enterprise.With the development of Chinese corporate bonds,the traditional rating method is no longer applicable.In recent years,decision tree,neural network,support vector machine and other machine learning algorithms have entered the field of credit rating,and achieved good results.In this paper,2016 listed companies issued corporate bonds as an example,the use of support vector machine algorithm for credit rating.And through the comparison of different kernel functions in the support vector machine,the most suitable kernel function is selected and it is foundthat the classification with the linear kernel function is better in the problem of bond credit rating.
作者
徐闪赏
Wang Zhou-wei;Fu Yi;Xu Shan-shang
出处
《金融管理研究》
2018年第2期63-82,共20页
The Journal of Finance and Management Research
关键词
公司债
信用评级
支持向量机
Corporate Bonds
Credit Rating
Support Vector Machine
