基于ENet-COX模型的中小企业信用风险评估研究

A Research on Credit Risk Assessment of Small and Medium-sized Enterprises Based on ENet-COX Model

中小企业的信用风险评估是解决中小企业融资难问题的前提。文章利用中小企业的财务指标和宏观经济指标建立评估指标体系,采用弹性网络和动态COX模型结合的方法来评估中小企业信用风险,再利用2004年6月至2020年6月上市中小企业的季度财务数据和宏观数据进行实证研究。结果表明,弹性网络变量选择模型兼顾了Ridge和Lasso的优点,能筛选出有效的信用风险评估指标,且ENet-COX的模型准确度高达86.63%,优于Ridge-COX模型和Lasso-COX模型。在加入宏观经济变量后,ENet-COX模型的准确度提高到87.21%,说明宏观经济因素对中小企业信用风险的影响不容忽视。

The credit risk assessment of SMEs is the key to solving the problem of financing difficulties for them.The paper uses the financial indicators and macroeconomic indicators of SMEs to establish an evaluation index system,and uses the combination of the elastic net and the dynamic COX models to assess the credit risk of small and medium-sized enterprises.Then,the empirical research is conducted by using quarterly data of listed SMEs from June 2004 to June 2020.The results show that the elastic net variable selection model takes into account the advantages of Ridge and Lasso,which can screen out effective credit risk assessment indicators,and the ENet-COX model’s accuracy is as high as 86.63%,much higher than that of Ridge-COX and Lasso-COX.After adding macroeconomic variables into it,the accuracy of the ENet-COX model goes up to 87.21%,indicating that the impact of macroeconomic factors on the credit risk of SMEs cannot be ignored.