基于神经网络的商业银行农业信贷风险评估研究

发布时间：2018-05-11 15:56

本文选题：BP神经网络 + 农业类上市企业　；参考：《湖南大学》2012年硕士论文

【摘要】：农业是国民经济的基础产业，在我国经济发展过程中发挥着重要作用。随着我国工业化和城市化进程的加快，农业在三大产业中的比重逐年下降，但是农业与其他产业的发展却更加紧密。作为一个弱势产业，农业需要通过工业化的发展为其提供支持，同时国家也在积极采取措施解决“三农”问题。农业经济的发展需要国家及金融机构部门加大对农业的资金投入力度，，从而使得农业信贷项目得到有效的支持和良好的发展。信贷业务是商业银行的一项传统业务，是目前乃至今后很长一段时间内商业银行的主要利润来源。然而近几年我国农业行业的信贷违约程度一直居全行业之首，作为重点扶持农业发展的商业银行而言，对农业信贷风险的预警、评估和防范是其面临的主要问题。如何构建适合我国现实情况的农业企业信贷风险评估模型，提高商业银行农业信贷风险管理水平，加强商业银行信贷决策，更好地防范和应对信贷风险是本文要解决的主要问题。早期的信贷评估方法比较容易受主观因素的影响，而神经网络模型在进行信贷风险评估过程中，不需要建立模型，可以将定性因素和定量因素综合考虑，将相关数据输入神经网络就能对数据之间的关系进行总结，并且神经网络对数据的处理有着良好的自适应性以及很强的学习、模仿、抗干扰能力，因此神经网络模型能够灵活地处理多变量的复杂环境，有效地表示出信贷指标和信用等级之间的非线性映射关系。本文在总结国内企业信贷风险评估文献的基础上，运用显著性分析和主成分分析构造出农业类上市企业的信贷风险评估指标，并将神经网络模型和Matlab软件工具相结合，提出了基于BP神经网络的农业企业信贷风险评估模型，并对模型进行调试，最终结果显示神经网络模型对农业上市企业信贷风险评估的准确率达到了85%，具有较高的精度，通过建立神经网络信贷风险评估模型从而为商业银行发放农业类贷款提供依据，达到规避农业企业信贷风险、降低银行不良贷款比率、减少银行经营成本的目的。
[Abstract]:Agriculture is the basic industry of national economy and plays an important role in the process of economic development of our country. With the acceleration of industrialization and urbanization in China, the proportion of agriculture in the three industries is decreasing year by year, but the development of agriculture and other industries is more closely. As a weak industry, agriculture needs to provide support through the development of industrialization. Meanwhile, the country is also actively taking measures to solve the "three rural" problems. The development of agricultural economy requires the government and financial institutions to increase the investment in agriculture, so that the agricultural credit project can be effectively supported and well developed. Credit is a traditional business of commercial banks, which is the main profit source of commercial banks for a long time. However, in recent years, the credit default degree of agricultural industry in China has been the first in the whole industry. As a commercial bank which focuses on supporting the development of agriculture, the assessment and prevention of agricultural credit risk is the main problem it faces. How to construct the agricultural enterprise credit risk assessment model suitable for the reality of our country, how to improve the agricultural credit risk management level of the commercial bank, strengthen the commercial bank credit decision-making, Better prevention and response to credit risk is the main problem to be solved in this paper. The early credit assessment methods are easily influenced by subjective factors, but the neural network model does not need to establish a model in the process of credit risk assessment, so it can take both qualitative and quantitative factors into account. The relationship between the data can be summarized by inputting the relevant data into the neural network, and the neural network has good adaptability and strong learning, imitation and anti-interference ability to deal with the data. So the neural network model can deal with the complex environment of multivariable flexibly and express the nonlinear mapping relationship between credit index and credit grade effectively. On the basis of summarizing the literature of credit risk assessment of domestic enterprises, this paper constructs the credit risk assessment index of agricultural listed enterprises by using significant analysis and principal component analysis, and combines the neural network model with Matlab software tools. The credit risk assessment model of agricultural enterprises based on BP neural network is put forward, and the model is debugged. The final results show that the accuracy of the neural network model for agricultural listed enterprises' credit risk assessment reaches 855.The model has a high accuracy. Through the establishment of neural network credit risk assessment model to provide the basis for commercial banks to issue agricultural loans to avoid the credit risk of agricultural enterprises reduce the ratio of non-performing loans of banks and reduce the operating costs of banks.
【学位授予单位】：湖南大学
【学位级别】：硕士
【学位授予年份】：2012
【分类号】：TP183;F832.43

【参考文献】