上市公司财务困境预测的改进马田系统方法

Improved Mahalanobis-Taguchi System Used to Predict Financial Distress of Listed Corporation

针对上市公司财务困境预测中产生的不平衡数据,研究了一种用于分类不平衡数据的改进的马田系统,并给出实施过程和分类器。对若干UCI数据集进行了实验比较,结果表明该方法对于不平衡数据有较好的分类效果,且能筛选出重要变量以降维。最后,运用该方法对上市公司的财务困境进行预测的实证研究结果表明,该方法在不平衡数据的条件下在识别处于财务困境的上市公司方面具有良好的效果。

According to the imbalanced data born in predicting the financial distress of listed corporation,this paper proposes an improved Mahalanobis-Taguchi system（MTS） used to classify the imbalanced data,and gives the implementation process and the classifier.Then it makes a experimental analysis on the UCI data set.The results indicate that the proposed method is very effective not only in the classification of imbalanced data but also in the reduction of dimensionality.The empirical research result on predicting the financial distress of listed corporation indicates that the improved MTS is more effective to identify the listed corporate in financial distress.