集成学习方法在企业财务危机预警中的应用  

Ensemble Learning Method and Its Application in Enterprise Financial Crisis Early Warning

作　　者：梁明江[1] 庄宇[1] 

机构地区：[1]西北工业大学管理学院,西安710129

出　　处：《软科学》2012年第4期114-117,共4页Soft Science

基　　金：国家社会科学基金一般资助项目(05BJY021)

摘　　要：以我国制造业上市公司为样本数据,用支持向量机作为基分类器的集成学习方法来预测企业的财务危机,通过具体实验分析可知:集成学习比单个基分类器的预测准确率提高了4个百分点,且稳定性更高,有效地提高了模型的预测精度,使得模型更具有准确性和应用性。基于支持向量机的集成学习方法在构建我国制造业上市公司财务危机预警模型上是有效的,且达到一定的财务危机预警效果。Taking listed manufacturing companies in China as the sample,the Ensemble Learning Method by which Support Vector Machine serves as base classifiers can be used to predict enterprise financial crisis.It can be seen from the experimental analysis that prediction accuracy of Ensemble Learning Method with higher stability has risen by four percent compared with single base classifier.The model presents higher prediction accuracy and applicability.It is valid for the Ensemble Learning Method based on Support Vector Machine to construct listed manufacturing company＇s financial crisis early warning model in China and the Ensemble Learning Method achieves certain effect from the perspective of enterprise financial early warning.

关 键 词：企业财务预警 集成学习方法 支持向量机 上市公司 

分 类 号：F275.5[经济管理—企业管理]