论文总字数：54190字

摘 要

中小企业融资难、融资贵等问题在我国以及世界其他国家存在已久。2018年，我国经济增速放缓、流动性趋于短缺，该问题更加突出。在这样的情况下，许多民营企业经营状况步履艰难。不论对商业银行还是政府来说，建立适合的中小企业信用评级模型变得十分重要，有助于企业有更多融资发展的机会。

本文采用了CCER经济金融数据库中的中小板块数据来进行实证研究。本文从中小企业的财务指标及其他定性指标出发，以资产负债率60%为高风险临界点，使用2016年到2018年的数据，对来自五个行业的500家中小企业以1:1:1:1:1的形式，在进行因子分析的前提下，将十大主成分因子带入模型中进行实证研究，使用logistic模型和神经网络模型来进行实证分析，通过OR值排列大小来找到对企业风险影响的因素的大小关系。通过实证研究发现这两个模型都能很好的预测中小企业的信用风险，且偿债能力指标与盈利能力指标对信用风险的影响最大，相反地，定性指标并不显著，特别是本科生占比人数没有很大的影响。同时五大行业中，批发和零售业的违约可能性更高一些。

本文最后对中小企业提升风险管理能力及银行建立科学的信用风险体系能提出了建议。

关键词：因子分析 logistic模型 OR值 神经网络

Credit Evaluation Model and empirical study for Small and Medium-sized Enterprises in commercial banks

Abstract

The problems of financing difficulties have existed both in China and other countries in the world for many years. In 2018, China's economic growth slowed down and liquidity tended to be short, which made the problem more prominent. Under such circumstances, many private enterprises are struggling. Whether for commercial banks or governments, it is very important to establish a suitable credit rating model for SMEs, which will help enterprises to have more opportunities for financing development.

We obtain enterprise credit information in CCER economic and financial database. This paper starts from the financial indicators and other qualitative indicators of small and medium-sized enterprises. With asset-liability ratio of 60% as the high-risk critical point, using data from 2016 to 2018, 500 small and medium-sized enterprises from five industries are processed in the form of 1:1:1:1:1, on the premise of factor analysis, ten principal component factors are introduced into the model for empirical research. Using logistic model and neural network model. Empirical analysis is conducted to find the relationship between factors affecting enterprise risk through the ranking size of OR value. Empirical research shows that these two models can predict the credit risk of SMEs very well, and Solvency Index and profitability index have the greatest impact on credit risk. On the contrary, the qualitative indicators showed insignificant especially the proportion of undergraduates. At the same time, wholesale and retail industries are more likely to default.

Finally, this essay offer advices for SMEs in order to improve their abilities of risk management and set scientific credit evaluation model.

Key words: factor analysis ; Logistic Model ;OR Value; Neural Network

目录

摘 要 I

Abstract II

第一章 引言 1

1.1 研究目的与意义 1

1.2相关研究背景 1

1.2.1国外相关研究回顾 1

1.2.2国内相关研究回顾 2

1.3创新点与研究方法 4

第二章 数据和方法论 5

2.1数据的说明 5

2.2研究思路与方法 8

2.2.1.样本企业行业分布表 8

2.2.2 因子分析理论 8

2.2.3 logistic模型理论 9

2.2.4 神经网络理论 10

第三章 中小企业信用风险实证研究 12

3.1因子分析 12

3.1.1因子模型检验 13

3.1.2公因子方差 13

3.1.3解释的总方差和碎石图 13

3.1.4因子载荷矩阵 14

3.1.5因子得分系数矩阵 16

3.2 logistic模型 16

3.3 神经网络 20

第四章 结论与建议 23

4.1模型检验结论 24

4.2相关投资及决策建议 24

结语 27

参考文献 28

致谢 29

附录 30

第一章 引言

1.1 研究目的与意义

请支付后下载全文，论文总字数：54190字