摘要
介绍了支持向量机的分类原理。通过72家上市企业的财务数据建立支持向量机的企业财务预警模型。并利用该模型对上市企业进行两类和三类模式识别。在此过程中还分析了采用不同核函数和参数的仿真结果,表明与传统的财务预警相比,SVM模型具有非线性和良好的泛化能力。
The support Vector machine （SVM） theory is introduced. A financial warning model of listed enterprises is established by the financial data of 72 listed enterprises. The SVM method is appled to two and three kind of pattern recognitions of the listed enterprises. In the experiment, the results are analyzed and compared by selecting different kernel functions and parameters. Compared with the traditional financial warning, the test shows the SVM model has the non-linearity and good generalization ability.
出处
《科学技术与工程》
2008年第5期1315-1318,共4页
Science Technology and Engineering
基金
江苏省自然科学基金(BK2007016)资助
关键词
财务预警
支持向量机
交叉验证
核函数
泛化
financial warning support vector machine cross validation kernel function generalization
