摘要
信用风险对一个上市公司来说十分关键,而信用风险受到多种因素的综合作用,变化十分复杂,当前信用风险预警方法无法反映其复杂的变化特点,使得信用风险预警错误率相当的高,信用风险预警结果不可靠。为了获得理想的信用风险预警效果,提出粒子群优化算法和支持向量机的上市公司信用风险预警方法。首先,分析上市公司信用风险预警原理,指出影响上市公司信用风险预警结果的重要因素;然后,将上市公司信用风险预警问题看作是一个多分类问题,通过支持向量机对上市公司信用风险变化特点进行深度分析和挖掘,建立上市公司信用风险预警分类器,并引入粒子群优化算法对上市公司信用风险预警分类器参数进行优化;最后,采用具体实例分析了上市公司信用风险预警效果。结果表明,文中方法的上市公司信用风险预警正确率超过90%,远远高于实际应用的85%,而且上市公司信用风险预警错误率要小于当前经典的上市公司信用风险预警方法,验证了所提方法的优越性。
Credit risk is very critical for a listed company. It is affected by a combination of factors and its changes are complex. However,the current credit risk early warning methods fail to reflect the characteristics of complex changes,which makes the error rate of credit risk early warning quite high and the result of credit risk early warning unreliable. In order to obtain an ideal credit risk early warning result,a listed companies′ credit risk early warning method based on particle swarm optimization(PSO)algorithm and support vector machine(SVM)is proposed. In the method,the principle of credit risk early warning of listed companies is analyzed,and the important factors that affect the results of credit risk early warning of listed companies are pointed out. And then,the credit risk early warning of listed companies is regarded as a difficulty in multi-classification. The SVM is used to deeply analyze and mine the change characteristics of listed companies′ credit risk to establish the credit risk early warning classifier of listed companies. Furthermore,the parameters of credit risk early warning classifier of listed companies are optimized by introducing PSO algorithm. Specific examples are employed to analyze the effect of credit risk early warning of listed companies. The results show that the accuracy of the credit risk early warning of listed companies,obtained with the proposed method,exceeds 90%,which is much higher than 85% existing in the practical application. The error rate of credit risk early warning of listed companies obtained with the proposed method is lower than that with the current classical credit risk early warning methods for listed companies,which verifies the superiority of the proposed method.
作者
周树功
李娟
ZHOU Shugong;LI Juan(Department of Mathematics and Information Sciences,Tangshan Normal University,Tangshan 063000,China)
出处
《现代电子技术》
北大核心
2020年第11期72-75,共4页
Modern Electronics Technique
基金
河北省社会科学发展研究课题(201804020101)。
关键词
信用风险
预警错误率
上市公司
多分类问题
支持向量机
预警正确率
credit risk
early warning error rate
listed company
multi-classification problem
SVM
early warning accuracy rate