节点文献

基于BP神经网络的商业银行客户信用风险评价研究

Research on the Credit Risk Evaluation of Commercial Bank Based on Neural Network

【作者】 邵海宏

【导师】 鞠晓峰;

【作者基本信息】 哈尔滨工业大学 , 技术经济及管理, 2007, 硕士

【摘要】 信用风险是金融机构面临的最主要风险,发达国家商业银行对信用风险管理比较成熟,在实践和理论上已形成相应体系,信用风险分析不断尝试采用新技术方法。我国商业银行信用风险管理体系不健全,管理技术较简单,还不能满足商业银行对信用风险管理要求。所以,研究信用风险评价对提高我国金融机构竞争能力,对商业银行信用风险管理有重要理论与现实意义。商业银行的发展对整体经济的发展有着重要作用。发达国家对信用风险评价研究已是成熟阶段,而我国正处于初步阶段,因此,本文首先提出研究的背景,概述了国内外研究现状并加以评析,找出其中的不足和明确本文应该研究的重点。其次,论证了信用风险评价对信用风险管理的重要性,商业银行信用风险相关概念的界定,信用风险和BP神经网络的理论基础,及构建信用风险评价指标体系的必要性。再次,借鉴国内外研究学者及标准普尔资信公司的研究,基于客户的系统性和盈利发展性构建了商业银行信用风险评价指标体系。评价体系包括企业素质、企业忠诚度、企业规模、偿债能力、盈利能力、发展能力六大类别,由偿还到期贷款情况、净资产收益率等26个二级指标组成。接着,基于BP神经网络法构建了商业银行信用风险评价模型,利用T检验和因子分析分析出贷款到期偿还率、总资产报酬率、净利润增长率等9个指标具有统计意义。最后,收集某商业银行161笔贷款业务进行实证分析,利用“3σ”评价规则确定客户初始信用水平,并运用Matlab软件对构建的模型进行检验。本文重点研究商业银行信用风险评价指标体系的构建,结合该体系基于BP神经网络构建了信用风险评价模型,并对模型进行了实证分析,从而为商业银行降低信用风险提供一定的借鉴意义。

【Abstract】 Credit risk is one of the most important risks that the financial institutions face. As the economic globalization process accelerated, credit risk management in the past 20 years has made tremendous progress. The credit risk management in developed commercial banks is more mature and more mature in practice and theory corresponding system. Credit risk analysis adopts new techniques day by day. By comparison, China commercial banks’credit risk management system is not perfect and management technology is relatively simple, it still can not meet the commercial banks’requirements to manage credit risk.The development of commercial bank plays an important role in the whole economy development. The credit risk management in developed commercial banks is more mature but premature in our country. So this paper starts from the research background of credit risk, studying and remarking on the research status of domestic and overseas’finding the drawback of present research as for which is the stressed point of this paper. Then this paper demonstrates the importance of credit risk evaluation to credit risk management、the definition to the related conceptions of commercial bank’s credit risk. What’s more, it demonstrate the theory foundation of credit risk and BP neutral network、the necessity of building credit risk evaluation system. Secondly, it adopts the overseas scholars’research result and S&P credit corporation. Basing on the systematic and profitable characteristics, it designs the credit risk evaluation index system of the commercial bank. The evaluation index system contains the enterprise’s aptitude, the enterprise loyalty、the enterprise scale, the debt liability, the profitability and the grow ability, and all these index is make up of second index, including pay back status, net asset’s yield and other 24 index. Basing on these indexes and BP neutral network method, the paper establishes the credit risk evaluation model of the commercial bank. Through the T-test and factor analysis method, we know that the nine indexes including debt payback period and ratio、the whole asset’s profit rate、net profit increase rate have statistical meaning. At last, the author collects 161 loan operation of one commercial bank and analyzes it by the methods of the above. It uses the 3sigma evaluation rule to determine the initial customer’s credit level and uses the Matlab to test the model. This article primarily studies how to build index system to evaluate credit risk of commercial bank, which is based on BP neutral network. The paper also use the index system to demonstrate a practical model. All of these is helpful for commercial bank to lower the credit risk.

【关键词】 信用风险风险评价BP神经网络
【Key words】 credit riskrisk evaluationBP neural network
  • 【分类号】F224;F832.4
  • 【被引频次】26
  • 【下载频次】1448
  • 攻读期成果
节点文献中: