基于数据挖掘的在信用卡客户风险管理及消费行为中的研究
发布时间：2018-07-24 22:01
【摘要】：随着商业银行信用卡业务的不断提高,数据挖掘技术开始应用于信用卡风险管理系统,该技术的运用存在着非常深远的意义。本文主要是研究了数据挖掘技术在信用卡风险管理系统里的应用。并且对该项技术的应用前景进行了预测。 在数据挖掘技术过程中主要通过SLIQ算法进行,SLIQ算法是一种基于决策树模型的模型,本文主要是运用决策树模型使其与具体业务程序相结合,从而在平时的客户审核操作中实现与实际状况相符合,这样就在很大程度上使客户分类以及信用评分可以更为合理,在充分保证算法科学合理的基础上,实现数据挖掘技术以及算法的优化。 对于系统集成,本研究细致深入的对主要的功能模块进行了阐明,同时简要概括了每一个功能模块的作用,并对该系统集成的体系架构以及相关技术进行了分析,在这里利用系统架构图的手段对其作出说明。另一方面,在实证部分本研究细致深入的阐明了某中小股份制商业银行的系统问的关联和系统的用户界面与业务流程逻辑。在对其中的关键技术的可行性以及功能进行验证的基础上,笔者还实现了其中的部分功能。在该部分笔者将描述统计、决策树模型以及利用层次分析法构建的评分表有机的结合在了一起,为信用卡客户进行定量的评分评价奠定了基础。同时对测试结果的准确程度进行了细致的验证,除此之外,笔者还分析了现阶段系统之中具有的不足之处与将来需要改进的方向。
[Abstract]:With the continuous improvement of credit card business in commercial banks, data mining technology has been applied to credit card risk management system. This paper mainly studies the application of data mining technology in credit card risk management system. The application prospect of this technology is forecasted. In the process of data mining, the SLIQ algorithm is a model based on the decision tree model. This paper mainly uses the decision tree model to combine it with the specific business program. In order to achieve in the usual customer audit operation and the actual situation is consistent, so that to a large extent, customer classification and credit scoring can be more reasonable, on the basis of fully ensuring the scientific and reasonable algorithm, Data mining techniques and algorithms are optimized. For the system integration, the main function modules are explained in detail, and the function of each function module is briefly summarized, and the architecture and related technologies of the system integration are analyzed. The system architecture diagram is used here to illustrate it. On the other hand, in the empirical part, the paper clarifies the connection of the system and the user interface and business process logic of a small and medium-sized joint-stock commercial bank. On the basis of verifying the feasibility and function of the key technology, the author also realizes some of the functions. In this part, the author combines the description statistics, the decision tree model and the score table constructed by AHP, which lays the foundation for the quantitative evaluation of credit card customers. At the same time, the accuracy of the test results is verified in detail. In addition, the author also analyzes the shortcomings of the present system and the direction of improvement in the future.
【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】：F832.2;F224;TP311.13
本文编号：2142787
[Abstract]:With the continuous improvement of credit card business in commercial banks, data mining technology has been applied to credit card risk management system. This paper mainly studies the application of data mining technology in credit card risk management system. The application prospect of this technology is forecasted. In the process of data mining, the SLIQ algorithm is a model based on the decision tree model. This paper mainly uses the decision tree model to combine it with the specific business program. In order to achieve in the usual customer audit operation and the actual situation is consistent, so that to a large extent, customer classification and credit scoring can be more reasonable, on the basis of fully ensuring the scientific and reasonable algorithm, Data mining techniques and algorithms are optimized. For the system integration, the main function modules are explained in detail, and the function of each function module is briefly summarized, and the architecture and related technologies of the system integration are analyzed. The system architecture diagram is used here to illustrate it. On the other hand, in the empirical part, the paper clarifies the connection of the system and the user interface and business process logic of a small and medium-sized joint-stock commercial bank. On the basis of verifying the feasibility and function of the key technology, the author also realizes some of the functions. In this part, the author combines the description statistics, the decision tree model and the score table constructed by AHP, which lays the foundation for the quantitative evaluation of credit card customers. At the same time, the accuracy of the test results is verified in detail. In addition, the author also analyzes the shortcomings of the present system and the direction of improvement in the future.
【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】：F832.2;F224;TP311.13
【引证文献】
相关博士学位论文 前1条
1 杜云生;信用卡消费市场细分研究[D];北京理工大学;2014年
，本文编号：2142787
本文链接：https://www.wllwen.com/guanlilunwen/bankxd/2142787.html
