基于Lasso+SVM的制造业上市公司财务风险组合预警模型

根据组合预测思想构建了基于Lasso+SVM的制造业上市公司财务风险组合预警模型,包括串联型组合和信息融合型组合两种,并选取22个财务指标建立了财务预警指标体系,对我国86家制造业上市公司的财务状况进行了预测,还与单一风险预警模型预测效果进行了比较,结果发现:财务风险组合预警模型的预测效果明显高于单一预警模型,用第t-1年的财务数据进行预测的准确率达到了95%以上;串联型组合预警模型的预测效果最优,用第t-1和t-2年的财务数据进行预测的准确率分别达到了100%和90%.

Lasso+SVM-based Early Warning Model of Financial Risks for Manufacturing Listedcompanies

Based on the combination forecasting idea,this paper builds a Lasso and SVM-based financial risk combination early warning model for manufacturing listed companies,including serial combination and information fusion combination.The paper selects 22 financial indicators as research variables to build a financial early warning indicator system.The financial status of 86 listed manufacturing companies in China was predicted,and the accuracy of these two models was given.The results were further compared with the prediction results of the single early warning model.It is significantly higher than the single early-warning model,and the accuracy of prediction using the data of year t-1 has reached more than 95% the prediction effect of the series-type combined early-warning model is the best.The accuracy rates of predictions using the t-1 and t-2 years data reached 100% and90% respectively.