基于数据包络分析：决策树的企业财务评价与预警--以西部四省上市企业为例

摘要

目前，对于企业财务状况评价与预警的研究主要依据ST与非ST两类企业进行分组，类别过于笼统，从而评价结果较为单一，难以衡量企业财务状况的变化过程。本文首先依据数据包络分析理论对上市公司的财务状况进行评估，将其划分为五个层次，进而引用C5.0决策树算法学习企业财务状况等级的划分规则，从而得到更加连续的企业财务评估方法，也适用于财务决策的检验与修正，使得对企业财务状况的研究更加贴近实际，评价与预警效果更加准确、实用。%In the studies of financial evaluation and warning at present, enterprises were divided into two groups according to ST or not. So the classes are too general and categories are not well defined, they could not reflect the process character from changes of company's financial situation. This paper uses DEA assessment to evaluate different enterprises' lever, according to financial data in western china. Then we differ sample sets with Decision tree algorithm to establish financial evaluation and warning system. It can provide a transitional evaluation and available for financial decisions' modification and measurement, make the financial assessment more accurate and practical.