上市公司会计信息失真识别研究——CART与MDA模型应用比较

通过分别运用分类回归树与多元判别分析模型,对196家上市公司年度财务报告数据进行建模,再利用1624家上市公司样本对所建模型进行识别能力检验.研究发现:分类回归树假设条件限制少,学习与识别能力强,对检验样本整体正确识别率达到81%以上,将第二类错误率有效控制在20%以下,留存收益在资产总额中的比重对会计信息失真识别具有很高的指示作用.

An Empirical Research on Identification of False Financial Statements:A Comparative Analysis of CART and MDA Models

In our research we employed CART(Classify and Regression Trees) and MDA(Multiple Discriminant Analysis) to establish two models with annual financial statements of China's 196 listed companies in 2006,and then we evaluated the model's accuracy with the data from 1624 listed companies in 2007.In the empirical study,we found that: CART model has a strong capacity for learning and identifying,which can identify more than 81% of the test samples correctly;CART is a convenient tool which has lower second-class m...