Research on Bond Default Risk of LN Co., Ltd.

本研究以我国首例国企可转债违约LN股份债券违约事件为切入点，运用机器学习方法构建企业债券违约预警模型。在初步选取财务指标的基础上，采用机器学习随机森林模型，自动选取财务指标特征、可转债市场特征与宏观经济变量等多维度特征的重要性，构建风险评分体系，训练与预测影响企业违约最具预测力的变量，捕捉风险波动趋势并提前预警，提高信用风险识别能力。This study takes the bond default event of LN Co., Ltd., the first state-owned enterprise convertible bond default in China, as the entry point, and uses machine learning methods to construct an early warning model for corporate bond defaults. On the basis of preliminary selection of financial indicators, a machine learning random forest model is adopted to automatically select the importance of multi-dimensional features such as financial indicator features, convertible bond market features, and macroeconomic variables, construct a risk scoring system, train and predict the variables with the most predictive power for corporate defaults, capture the trend of risk fluctuations, and provide early warnings, so as to improve the ability to identify credit risks.