CAJ | 학술논문

在Bagging分类器集成算法的基础上，利用马田系统实现特征选择，提出MTS-Bagging分类器集成方法，并对4个UCI数据集进行数值分析。将MTS-Bagging神经网络集成方法应用于企业财务危机预测，利用2003--2008年我国330家上市公司的财务数据建立企业危机预测模型。实证结果表明，利用MTS-Bagging神经网络集成方法对财务危机的预测精度不仅优于单一分类器，也优于Bagging集成算法。
재Bagging분류기집성산법적기출상，이용마전계통실현특정선택，제출MTS-Bagging분류기집성방법，병대4개UCI수거집진행수치분석。장MTS-Bagging신경망락집성방법응용우기업재무위궤예측，이용2003--2008년아국330가상시공사적재무수거건립기업위궤예측모형。실증결과표명，이용MTS-Bagging신경망락집성방법대재무위궤적예측정도불부우우단일분류기，야우우Bagging집성산법。
On the basis of Bagging classifier ensemble algorithm, this paper proposes a new classifier ensemble MTS-Bagging technology through using Mahalanobis-Taguchi system for feature selection,and studies 4 UCI data sets in the data experiments. And it applies the technology of MTS-Bagging neural network ensemble to predict financial crisis,and uses the financial data of 330 China＇s listed companies from 2003 to 2008 to build the model for predicting enterprise＇s financial crisis. The experiment results show that the precision of MTS-Bagging neural network ensemble predicting enterprise＇s financial crisis excels that of single classifier,and also excels that of bagging algorithm.