基于多角度的银行绿色信贷风险评价研究

发布时间：2018-07-04 08:23

本文选题：层次分析法 + BP神经网络　；参考：《延边大学》2017年硕士论文

【摘要】：随着我国信用风险研究的发展,人类在享受成果的同时,绿色信贷风险问题也日渐突出。本文依托大数据分析理论对绿色信贷这一热点问题,结合所学知识,运用统计分析中的层次聚类分析、主成分分析等特征工程方法进行研究,同时利用机器学习中的监督学习式BP神经网络模型进行仿真分析,进而得到更加科学的结果。银行机构在企业贷款中最重要的风险之一就是信贷风险,实行风险防控对机构的长远发展有至关重要的作用。本文所阐述的绿色信贷是指银行等金融机构在进行企业信贷时结合生态保护建设与开发等方面,在原有信用风险基础之上,重新评定企业的信贷能力。纵观文献可了解到国内外诸多学者对信用风险有着大量研究,但针对绿色信贷这一刚刚兴起的热点问题,还处在探索阶段,发展还不成熟。本文探索性构建了两个数据挖掘模型,分别是基于AHP的绿色环保评级模型以及基于PCA的BP神经网路绿色信贷信用评估风险模型。首先,针对绿色信贷在我国处于起步阶段,缺乏相关数据,本文查阅相关文献并进行研究,结合银行现有信用风险体系,利用AHP构建绿色信贷信用风险评估指标层次体系。接着,通过东北证券同花顺提取2000多家上市公司基础数据,选取企业类型为综合型企业作为样本数据,应用系统聚类分析针对原有风险评价研究的10个指标进行分析,剔除高危风险企业。针对筛选后的企业增加4个绿色风险指标,以经济以及绿色风险指标作为一级指标,下设14个二级指标,其中10个为基础财务指标,4个为层次分析法准则层指标,利用主成分分析方法来构建模型与处理数据,通过PCA对14个指标进行处理,得出6个公共因子。最后利用BP神经网络进行仿真分析(matlab2012a),验证其可行性。以定量与定性分析的方式为绿色信贷风险评估的方法选择与运用提供参考。
[Abstract]:With the development of credit risk research in China, the problem of green credit risk is becoming more and more prominent. Based on the theory of big data analysis, this paper studies the hot issue of green credit, combining with the knowledge learned, using hierarchical cluster analysis, principal component analysis and other characteristic engineering methods in statistical analysis. At the same time, the supervised learning BP neural network model in machine learning is used for simulation and analysis, and more scientific results are obtained. The credit risk is one of the most important risks in the enterprise loan. The risk prevention and control plays an important role in the long-term development of the institution. The green credit described in this paper means that banks and other financial institutions reassess the credit ability of enterprises on the basis of the original credit risk in combination with the ecological protection construction and development of enterprise credit. Throughout the literature, we can find that many scholars at home and abroad have a lot of research on credit risk, but the green credit is still in the exploration stage, and the development is not mature. In this paper, two data mining models are constructed, one is based on AHP and the other is the risk model of credit assessment based on BP neural network based on PCA. First of all, in view of the green credit in our country is in the initial stage, the lack of relevant data, this paper refer to the relevant literature and research, combined with the existing credit risk system of banks, using AHP to construct the evaluation index system of green credit risk. Then, the basic data of more than 2000 listed companies are extracted through Tonghuashun of Northeast Securities, and the comprehensive enterprise is selected as the sample data. The system cluster analysis is applied to analyze the 10 indexes of the original risk evaluation research. Eliminate high-risk enterprises. In view of the increase of 4 green risk indicators in the selected enterprises, taking the economic and green risk indicators as the first class index, there are 14 secondary indicators, of which 10 are the basic financial indicators, 4 are the Analytic hierarchy process (AHP) criteria. The principal component analysis (PCA) method was used to construct the model and process the data. The 14 indexes were processed by PCA and 6 common factors were obtained. Finally, BP neural network is used for simulation analysis (matlab2012a) to verify its feasibility. It provides a reference for the selection and application of green credit risk assessment methods by quantitative and qualitative analysis.
【学位授予单位】：延边大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：O211.67

【相似文献】