摘要
本文利用相关财务危机理论建立影响上市公司财务危机的指标体系,通过粗集理论对这些指标进行约简获得核心指标,再利用支持向量机对核心指标建模得到企业财务危机预警模型,并运用到未来三年的财务危机预测当中。实证分析表明,本模型前两年的综合预警准确率达90%以上,证明了该模型有较强的预测能力。从财务危机预警结果来看,与传统SVM方法相比,粗集及遗传算法的引入不仅能够提高预警效率,而且能够提高预测精度,与实际企业财务情况基本一致。实际应用表明,在企业财务危机预警建模中,粗集理论的约简和遗传支持向量机方法的实施充分利用了样本数据本身特点,并为后续的优异预警结果提供良好地理论基础。
This paper builds the index system affecting financial crisis of listed companies with related theory, and re- duces these indicators by rough set theory to get core indicators, then it uses SVM to model the core indicators to get the model of enterprise financial crisis warning, and uses it to predict the financial crisis in the next three years. The finan- cial crisis warning results show, compared with the traditional SVM method, the introduction of rough sets and genetic al- gorithms not only can improve the efficiency of early warning, but also can improve the prediction accuracy, and this warning results are consistent with the actual financial situation of enterprises. Practical application shows that the enter- prise financial crisis modeling, reduction and implementation support vector machine genetic rough set theory take full advantage of the characteristics of the sample data itself, which provides good theoretical basis for subsequent warning ex- cellent results.
出处
《商业研究》
CSSCI
北大核心
2015年第6期104-113,共10页
Commercial Research
基金
国家自然科学基金项目
项目编号:61272506
关键词
制造业上市公司
财务危机预警
粗集
遗传算法
支持向量机
manufacturing industry listed companies
financial crisis warning
rough set
genetic algorithm
SupportVector Machine （SVM）
