基于BP神经网络的钢铁行业上市公司财务风险预警研究

发布时间：2018-10-15 11:11

【摘要】：钢铁行业是我国国民经济的支柱产业，，目前全球经济持续放缓，铁矿石、能源价格高位运行，使得原本处于产能过剩的我国钢铁行业经营状况进一步恶化，利润率大幅下滑，财务风险加剧。钢铁业一旦发生财务风险，不仅危及其自身的生存和发展，也会给投资者和其他关联产业带来损失。因此，构建一个有效实用的钢铁业上市公司财务风险预警模型，满足利益相关者的需要，具有较大的现实意义。本文从智能理论角度着手，首先对粗糙集理论及其约简属性和BP神经网络的基本工作原理做了介绍，提出一种将粗糙集与BP神经网络相结合的技术方法，把该方法应用于我国钢铁业上市公司财务风险预警研究中。首先，介绍了我国钢铁业上市公司财务风险预警研究的研究背景、研究意义和国内外研究现状，并指出了以前研究的成果及其实用性，论证了本次研究的必要性；其次，对财务风险做了界定，并对其形成因素做了详细分析，分别阐述了粗糙集理论和BP神经网络的基本工作原理，详细分析了二者结合的互补优势特性，为后文预警模型的建立打下了理论基础；再次，介绍了目前我国钢铁业上市公司财务风险的表现形式，并对影响风险产生的外部与内部因素做了全面分析，在该分析基础上结合钢铁行业自身特点，选取能够表现钢铁企业财务状况的财务指标和非财务指标，构建财务风险预警指标体系；最后，选取30家钢铁业上市公司为研究样本，对样本的指标数据按照上述方法进行处理，针对传统方法在预警模型建立方面存在的局限性，本文创造性地利用层次聚类分析将样本企业财务状况划分为递进的五级层次，构建BP神经网络预警模型，多级分类的财务状况为预警模型提供精确的输出层目标。训练BP神经网络后，用检测样本对其进行检验，证明模型预警效果良好。实验结果证实针对钢铁业上市公司所构建的粗糙集—BP神经网络财务预警模型是有效的。
[Abstract]:The iron and steel industry is the pillar industry of our national economy. At present, the global economy continues to slow down, and the iron ore and energy prices are running high. This has further worsened the operating situation of the steel industry in China, which was originally under overcapacity, and its profit margin has fallen sharply. The financial risk intensifies. Once the financial risk occurs in the steel industry, it not only endangers its own survival and development, but also brings losses to investors and other related industries. Therefore, it is of great practical significance to construct an effective and practical financial risk early warning model for listed steel companies to meet the needs of stakeholders. In this paper, from the angle of intelligence theory, the rough set theory and its reduction attribute and the basic working principle of BP neural network are introduced, and a technical method combining rough set with BP neural network is proposed. This method is applied to the financial risk early warning research of listed steel companies in China. First of all, it introduces the research background, significance and current situation of financial risk early warning research of listed steel companies in China, and points out the results of previous research and its practicability, and demonstrates the necessity of this study. Secondly, financial risk is defined, and its forming factors are analyzed in detail. The basic working principles of rough set theory and BP neural network are expounded respectively, and the complementary advantages of the two combination are analyzed in detail. It lays a theoretical foundation for the establishment of the early warning model. Thirdly, it introduces the expression forms of financial risks of listed steel companies in China, and makes a comprehensive analysis of the external and internal factors that affect the emergence of the risks. On the basis of this analysis, combining the characteristics of iron and steel industry, select the financial indicators and non-financial indicators that can express the financial situation of iron and steel enterprises, and construct the financial risk warning index system. 30 listed steel companies are selected as the research samples, and the index data of the samples are processed in accordance with the above methods, aiming at the limitations of the traditional methods in the establishment of early warning models. This paper creatively uses hierarchical clustering analysis to divide the financial situation of sample enterprises into progressive five levels, and constructs a BP neural network early warning model. The financial situation of multilevel classification provides the accurate output level target for the early warning model. After training BP neural network, it is proved that the early warning effect of the model is good. The experimental results show that the rough set-BP neural network financial early warning model for steel listed companies is effective.
【学位授予单位】：燕山大学
【学位级别】：硕士
【学位授予年份】：2013
【分类号】：F275;F426.31;TP18

【参考文献】