节点文献

基于组合BP神经网络的企业预警模型实证研究

Financial Early Warning System Study Based on Combination BP Network

【作者】 孙毅

【导师】 张华伦;

【作者基本信息】 西安理工大学 , 企业管理, 2006, 硕士

【摘要】 竞争激烈的市场经济孕育发展机遇的同时,也暗藏着无尽的风险和危机。对于上市公司而言,因财务危机沦为“ST”板块甚至被迫退市的情况愈演愈烈。公司陷入财务危机,不仅危及自身的生存和发展,也给投资者、债权人带来巨大的损失。因此,构建一个有效实用的财务危机预警模型,满足利益相关者日益迫切的需要,己不单单是一个学术问题,而且成为影响我国资本市场健康发展的重要因素,具有很重要的现实意义。本文以制造类企业为研究对象,旨在解决当前传统财务危机预警指标导致的财务预警失效,并在吸收前人成果的基础上,提出了现金制预警指标,之后进行了模型构造和实证研究,对于完善我国企业财务危机预警方面的研究具有一定参考价值。首先,本文总结了国内外财务危机预警的研究现状,并对财务危机预警的相关理论作出了分析,尤其是财务危机的界定及其形成原因。其次,本文根据预警指标建立的意义和原则,提出了基于现金流量的10个财务指标,并通过对其进行T检验,验证其合理性,同时还给出了本文的预警准则定义。然后,本文在介绍了部分常用的财务预警分析方法之后,提出了基于主成分分析法和BP(反向传播,BackPropagation)神经网络的组合预测方法,并给出了具体的原理和算法,构建了财务危机的预警模型。在选择了152家上市公司作为实证分析对象之后,进行了实证研究。结果表明:组合预测模型集合了统计模型和人工智能模型的优点,在上市公司财务危机发生前两年有83.92%的判别准确率,具有较强的优越性和应用价值。最后,本文从企业的现金流、内部控制和经营战略三方面阐述了企业财务危机的管理对策,从而为企业在完成利用预警工具进行事前控制之后,采取措施弥补企业管理体系上的漏洞,减少企业发生财务危机的风险,提供了可借鉴的措施和方法。

【Abstract】 When fierce competition accompany with the opportunity, the market nowadays is more risk than any other days. So, as for the public companies, they are now facing a critical problem of being financial disabled, in other means is that they will have financial crisis or even worse being forced to quit list. If a company runs into such situation, it will surely bring threat to the existence and development of that company, and what’s more brings enormous loss to investors, creditors as well. So to establish a effective and practical early warning model on financial crisis, in order to meet the increasingly urgent needs of relevant interests party is not only an academic problem, but also becomes the important factor influencing sound development of capital market of our country, and has very important realistic meanings.In order to solve the existing problems that lies in the current financial early warning indexes and that the invalidation caused by financial information distortion, the thesis made a series of cash flow prediction indexes regarding manufacturing industry company as the research object. The thesis also made out a demonstration that supplied references to the study of financial early warning in our country. Firstly, by way of summarizing the current study situation about early warning of the financial crisis, the thesis analyzed the theory and the method of financial crisis early warning.Secondly, after giving the meaning and standard of our financial early warning system, the thesis creates ten indexes both based on cash flow, then we have them both under the T examination to prove their capability of doing financial early warning, meanwhile we give the identify of the early warning’s standard of this thesis.Thirdly, after introduce some common methods for analysis financial crisis; the thesis formed a new combination forecast model using combination forecast theory to combine the BP(Back Propagation)neural network with the factor analysis. The thesis use 152 public companies as the study objects, after doing the experiment the result shows that this combination forecast model has the advantages of both tradition statistics model and artificial intelligence model, and the accurate radio is 83.92%, that means it will have great value in practical use.Finally, the thesis give some advises from three different aspects such as cash flow management, internal control and business strategy, to reduce the risk of being financial crisis, after having financial early warninganalysis. Thus this thesis provided a new way to the financial early warning for the manufacturing industry company.

  • 【分类号】F275;F224
  • 【被引频次】5
  • 【下载频次】609
节点文献中: 

本文链接的文献网络图示:

本文的引文网络