基于SVM算法的债券违约风险预测模型研究

Research on bond default risk prediction model based on SVM algorithm

近年来,我国债券市场交易规模不断扩大,在利益与风险并存的背景下债券违约发生的可能性随之上升,债券违约风险预测具有重要意义。通过收集2014-2020年发生债券违约公司的相关财务指标数据作为样本,利用支持向量机算法构建债券违约风险预测模型。研究结果显示,为保证模型的拟合和泛化能力,择优选择样本数据的划分比例和核函数,最终构建的支持向量机债券违约风险预测模型能够取得较高的正确率和特异度。该研究在上市公司债券违约风险预测模型构建、违约主体识别方面具有一定参考价值。

In recent years,the scale of bond market trading in China has continued to expand,and the possibility of bond default has increased with the coexistence of interests and risks.The risk prediction of bond default is of great significance.By collecting data on relevant financial indicators of companies that defaulted on bonds in 2014-2020 as a sample,a bond default risk prediction model is constructed using the support vector machine algorithm.The proportion of the sample data and the nuclear function are selected on merit to ensure the fit and generalization ability of the model,and the final model of bond default risk prediction based on support vector machine can obtain high accuracy and specificity.The research has some reference value in the construction of the bond default risk prediction model and the identification of default subject of listed companies.