商业银行个人住房贷款信用风险评估研究

发布时间：2018-01-21 00:47

本文关键词： 个人住房贷款信用风险评估主成分分析 RBF神经网络　出处：《首都经济贸易大学》2013年硕士论文　论文类型：学位论文

【摘要】：目前我国商业银行的个人住房贷款业务正处于快速发展的阶段，在发展过程中同时存在着诸如信用风险、市场风险、法律政策风险等各种类型的风险。商业银行将信用风险作为风险控制管理关注的重点，因此建立一个科学全面的个人住房贷款信用风险评估模型对银行个人贷款业务的稳步发展和风险防控工作是非常必要的。我国个人信用风险评估工作起步较晚，各家商业银行建立的评估体系尚未有统一的标准，且还存在一些不足，这些都制约了个人消费信贷业务的健康发展。本文通过对我国商业银行现行个人信用风险评估的分析，在借鉴西方发达国家和地区相对成熟体系经验的同时，充分考虑我国现实国情，得出一个相对合理全面的个人信贷信用风险评估指标体系。体系中加入了贷款状况指标，更贴近住房贷款业务，体现了专业性质，同时将资产负债状况分层细化考虑，体现住房、汽车、债券等各类型资产负债情况。本文在确定了信用风险评估指标的基础上，将运用基于主成分分析法的RBF径向基函数神经网络模型对信用风险进行预测评估，验证评估指标的现实效果。通过主成分分析法和RBF径向基函数神经网络的融合，可以对RBF径向基函数神经网络的输入变量进行筛选和降维，将累计贡献率大的变量集作为最终网络的输入变量。在此基础上对样本数据进行训练、测试，，通过实例数据证明取得了较好的评估预测效果，能够为商业银行的信用风险管理工作提供有效的支持。
[Abstract]:At present, the personal housing loan business of commercial banks in our country is in the stage of rapid development, such as credit risk and market risk in the process of development at the same time. Commercial banks take credit risk as the focus of risk control management. Therefore, it is necessary to establish a scientific and comprehensive personal housing loan credit risk assessment model for the steady development of bank personal loan business and risk prevention and control work. The evaluation system established by various commercial banks has not yet had a unified standard, and there are still some shortcomings, which restrict the healthy development of personal consumption credit business. Through the analysis of the current personal credit risk assessment of commercial banks in China, this paper takes full account of the actual situation of our country while learning from the relative mature system experience of western developed countries and regions. Get a relatively reasonable and comprehensive personal credit risk evaluation index system. The system adds loan status indicators, closer to the housing loan business, reflects the professional nature. At the same time, the state of assets and liabilities are classified and considered to reflect the housing, automobile, bonds and other types of assets and liabilities. On the basis of determining the credit risk evaluation index, this paper uses the RBF radial basis function neural network model based on principal component analysis to predict and evaluate the credit risk. Through the fusion of principal component analysis and RBF radial basis function neural network, the input variables of RBF radial basis function neural network can be screened and reduced. The variable set with large cumulative contribution rate is taken as the input variable of the final network. On this basis, the sample data are trained and tested, and good evaluation and prediction results are obtained by the example data. To provide effective support for the credit risk management of commercial banks.
【学位授予单位】：首都经济贸易大学
【学位级别】：硕士
【学位授予年份】：2013
【分类号】：F832.45

【参考文献】