基于费米分布函数的财务预警模型实证研究——以我国服务性行业上市公司为例

发布时间：2018-06-04 15:31

  本文选题：费米分布模型 + 因子分析　； 参考：《会计之友》2017年03期

【摘要】：通过引入一种具有统计学原理的费米分布模型对企业财务状况进行了预警实证研究,结果发现:当赋予费米分布模型在财务预警领域的物理内涵后,一定程度上能够对输入的企业综合财务得分值E进行准确预测,其准确性主要决定于研究样本费米面E_F的选择。E_F越接近样本的实际值,则预警准确率会显著提高;同时,获得具有正、负相关性的E值对该模型的财务预警准确率至关重要。采用因子分析法和正、负相关性财务指标算术和的方法,分别对获取的E值输入费米分布模型进行研究,表明采用因子分析法所得的E值由于考虑的企业财务指标过多过杂,对因子分析法所建模型和费米分布模型的预警准确率均产生了一定干扰。相比之下,采用正、负相关性财务指标算术和得到的E值能够有效提高费米分布模型的二进制预警准确率。
[Abstract]:By introducing a kind of Fermi distribution model with statistical principle, this paper makes an empirical study on the financial situation of an enterprise. The results show that: when the Fermi distribution model is endowed with the physical connotation in the field of financial early warning, To a certain extent, it is possible to accurately predict the integrated financial score E of the input enterprise. Its accuracy is mainly determined by studying the selection of sample Feimian ESP. The closer the Estack F is to the actual value of the sample, the higher the early warning accuracy rate will be significantly improved; at the same time, Obtaining E value with positive and negative correlation is very important to the financial early warning accuracy of the model. Using the methods of factor analysis and arithmetic sum of positive and negative correlation financial indexes, the input Fermi distribution model of the obtained E value is studied respectively. The results show that the E value obtained by factor analysis method is too complex because of the excessive financial indexes considered. The early warning accuracy of the model built by factor analysis method and Fermi distribution model is disturbed to a certain extent. By contrast, the arithmetic and E value of positive and negative correlation financial indexes can effectively improve the accuracy of binary early warning of Fermi distribution model.
【作者单位】： 哈尔滨理工大学经济学院;
【分类号】：F719;F715.5
，

本文编号：1977888