基于决策树算法的房贷信用风险评估研究

发布时间：2018-03-16 07:15

本文选题：房贷　切入点：信用风险评估　出处：《哈尔滨理工大学》2012年硕士论文　论文类型：学位论文

【摘要】：2008年美国次贷危机的爆发,使房贷信用风险引起全世界的关注。根据巴塞尔委员会提出的《新巴塞尔资本协议》的要求,金融市场风险或信用风险等单一风险已经不再是影响银行业经营困难的主要因素,而是多种因素(如：金融市场风险、信用风险、操作风险等)共同作用的结果。然而我国银行风险评估尚处于初级阶段,尤其对个人信用风险评估、管理方面的经验和方法都较为缺乏。基于此现状,本文以银行客户数据为依据对个人信用风险进行评估,为完善银行监管体系具有现实意义。本文开篇论述了国内外信用风险评估的研究现状以及研究成果,简单介绍了决策树算法和房贷信用风险,在分析房贷信用风险评估现状的基础上,进一步分析了房贷信用风险评估中存在的问题。对影响房贷信用风险评估的因素进行了分析,并设计出相应的评估指标,给出评估指标筛选的具体步骤,最终确定影响比较大的十个指标。紧接着通过对房贷信用风险评估方法的比较分析,选择了决策树算法来对房贷的信用风险进行评估,阐述了决策树算法所依托的原理和评估模型。根据以上对评估指标的设计以及评估方法的选择,选取了A银行为研究对象,对其房贷信用风险进行了评估,并根据评估结果提出了降低A银行房贷信用风险的策略。本文所采用的决策树算法能够准确的评估出房贷客户的信用等级,既适用于银行对老客户信用的跟踪评估,也适用于对新客户信用等级的预测,其评估结果可以成为银行为客户提供贷款的决策依据,能够为降低银行的房贷信用风险发挥巨大作用,保障银行业健康平稳的发展。
[Abstract]:In 2008, with the outbreak of the subprime mortgage crisis in the United States, the mortgage credit risk attracted worldwide attention. According to the request of the Basel Committee of the New Basel Capital Accord, The single risk, such as financial market risk or credit risk, is no longer the main factor that affects the banking management difficulty, but a variety of factors (such as: financial market risk, credit risk, etc.). However, the bank risk assessment in China is still in the initial stage, especially for the personal credit risk assessment, management experience and methods are relatively lacking. Based on bank customer data, this paper evaluates personal credit risk, which has practical significance for perfecting bank supervision system. At the beginning of this paper, the research status and achievements of credit risk assessment at home and abroad are discussed, and the decision tree algorithm and mortgage credit risk are briefly introduced. On the basis of analyzing the present situation of mortgage credit risk assessment, This paper further analyzes the problems existing in the assessment of mortgage credit risk, analyzes the factors that affect the assessment of mortgage credit risk, designs the corresponding evaluation index, and gives the concrete steps for the screening of the evaluation index. Then, through the comparative analysis of the methods of assessing the credit risk of housing loans, the decision tree algorithm is chosen to evaluate the credit risk of housing loans. This paper expounds the principle and evaluation model of decision tree algorithm. According to the design of evaluation index and the selection of evaluation method, Bank A is selected as the research object, and the credit risk of housing loan is evaluated. According to the evaluation results, the paper puts forward some strategies to reduce the credit risk of A bank. The decision tree algorithm used in this paper can accurately evaluate the credit rating of mortgage customers, which is not only suitable for the bank to track and evaluate the credit of the old customers, but also for the prediction of the credit rating of the new customers. The evaluation results can be used as the basis for banks to provide loans to customers, and can play a great role in reducing the risk of mortgage credit of banks, and ensure the healthy and stable development of banks.
【学位授予单位】：哈尔滨理工大学
【学位级别】：硕士
【学位授予年份】：2012
【分类号】：F832.45;F224

【参考文献】