基于决策树的银行监管分析系统的设计与实现
|
摘要
分类是数据挖掘的重要内容之一,在许多领域得到广泛应用,目前己有多种分类方法,其中决策树分类法在海量数据环境中应用最为广泛。
由于银行在电子化的发展过程中,积累了大量的客户信息,账户信息和交易信息,迫切需要银行监管部门提供更好的技术手段进行深层次的分析,挖掘客户价值,提示银行经营效益和风险。如何为银行业监管、决策提供科学依据,如何更好地为银行提供差异化服务依赖于银行监管手段提高。
本文首先针对银行业务领域数据挖掘的关键技术—决策树进行分析;然后结合银行业的具体情况,探讨了建立基于决策树分析方法的银行监管分析系统的可行性,并结合银行风险监管实际,构造了不良贷款分析决策树及其七条骨干树枝,设计实现了银行监管分析系统,并以分析查询子系统的实现为例,介绍了决策树分析方法的成功应用及预测结果分析,为决策者提供服务。本系统以现场检查操作流程为主线,采用模块化设计手段,直接提取金融机构业务数据、对相关数据进行分析树构建,从而使监管人员能直观判断、分析银行风险隐患,达到预警提示和有效分析的目的。本系统还针对银行业的应用特点,重点解决了系统的时间效率、空间效率及系统安全问题。
Classification is an important problem in data mining. Classification now hasbeen successfully applied to wide range of application areas. Many differenttechniques have been proposed for classification, decision tree classifiers have foundthe widest applicability in large_scale data mining enviroments.
Because in the process of electronical development for the bank, it alreadyaccumulated a great deal of customer's information, the account information and thetransaction information ,this urgently needs the banking management to provide thebetter analysis technique carries on an even profounder level, Mining the value of thecustomer , raising the business benefit. How to provide scientific basis for the bankingmanagement and decisions, how to provide the difference service for the bank aredepend on the method to raises of banking management analysis.
This thesis first analysis decision tree ,the key technique of the data mining forthe bank business;Then approach the feasibility of to establish the bankingmanagement analysis system based on data mining technology of decision tree,combine the actual circumstance of the banking. Combining the issue in thesupervisions of banks risk problems, the system constructed the decision-making treeof non-performing loans and its seven backbone sticks. Design and achieve thebanking management analysis system, take achieve of subsystem of analyze andinquire as example, introduce the success apply of technology of decision tree andanalyzes calculate result, pvovide the serve for policymaker. The operational processof spot cheeks as the main line, using modular design means, this system direct datefrom financial institutions business and analyzes relevant date. Then achieve thepurpose of warning suggestion and effective. Aiming at the applied characteristics ofthe banking, this system resolved emphases the time efficiency, the space efficiencyand systems safe problem.
由于银行在电子化的发展过程中,积累了大量的客户信息,账户信息和交易信息,迫切需要银行监管部门提供更好的技术手段进行深层次的分析,挖掘客户价值,提示银行经营效益和风险。如何为银行业监管、决策提供科学依据,如何更好地为银行提供差异化服务依赖于银行监管手段提高。
本文首先针对银行业务领域数据挖掘的关键技术—决策树进行分析;然后结合银行业的具体情况,探讨了建立基于决策树分析方法的银行监管分析系统的可行性,并结合银行风险监管实际,构造了不良贷款分析决策树及其七条骨干树枝,设计实现了银行监管分析系统,并以分析查询子系统的实现为例,介绍了决策树分析方法的成功应用及预测结果分析,为决策者提供服务。本系统以现场检查操作流程为主线,采用模块化设计手段,直接提取金融机构业务数据、对相关数据进行分析树构建,从而使监管人员能直观判断、分析银行风险隐患,达到预警提示和有效分析的目的。本系统还针对银行业的应用特点,重点解决了系统的时间效率、空间效率及系统安全问题。
Classification is an important problem in data mining. Classification now hasbeen successfully applied to wide range of application areas. Many differenttechniques have been proposed for classification, decision tree classifiers have foundthe widest applicability in large_scale data mining enviroments.
Because in the process of electronical development for the bank, it alreadyaccumulated a great deal of customer's information, the account information and thetransaction information ,this urgently needs the banking management to provide thebetter analysis technique carries on an even profounder level, Mining the value of thecustomer , raising the business benefit. How to provide scientific basis for the bankingmanagement and decisions, how to provide the difference service for the bank aredepend on the method to raises of banking management analysis.
This thesis first analysis decision tree ,the key technique of the data mining forthe bank business;Then approach the feasibility of to establish the bankingmanagement analysis system based on data mining technology of decision tree,combine the actual circumstance of the banking. Combining the issue in thesupervisions of banks risk problems, the system constructed the decision-making treeof non-performing loans and its seven backbone sticks. Design and achieve thebanking management analysis system, take achieve of subsystem of analyze andinquire as example, introduce the success apply of technology of decision tree andanalyzes calculate result, pvovide the serve for policymaker. The operational processof spot cheeks as the main line, using modular design means, this system direct datefrom financial institutions business and analyzes relevant date. Then achieve thepurpose of warning suggestion and effective. Aiming at the applied characteristics ofthe banking, this system resolved emphases the time efficiency, the space efficiencyand systems safe problem.
引文
[1] http://industry.ccidnet.com/art/465/20020517/14312_1.html CRM：撑起金融一片天.忻学庆. 2002 Vol.2 No.8
[2]http://www.chonna.com/knowledge/news_article_display.asp?depid=2&subdepid=3&articleid=297淘金：数据挖掘在CRM中的运用
[3]王向星，袁胜，刘笑东，基于数据仓库的CRM在银行业中的应用，计算机世界.2003 Vol.28 No.16
[4]葛晨霞，商业银行CRM的需求分析和系统实现，中国金融电脑，2003 Vol.12No.9
[5]周意，李峰峰，基于CRM的银行业务应用模块分析，中国金融电脑，2004 Vol.9No.4
[6]高秩新，实现更具数据驱动力的银行客户关系管理，中国金融电脑，2005 Vol.6No.70
[7] Harjinder S.GILL等著,王仲谋,刘书丹译，数据仓库-客户服务器计算指南，北京：清华大学出版社,西蒙与舒斯特国际出版公司，1997 Vol.1
[8]陈文伟，黄金才，数据仓库与数据挖掘，北京：人民邮电出版社2004 Vol.1
[9]李子木,莫倩,周兴铭，数据库技术的研究现状及未来方向，计算机科学，1998Vol.25 No.4
[10] http://www.creawor.com/biforum/bi_03.htm，数据挖掘的研究现状
[11]王珊等，数据仓库技术与联机分析处理，科学出版社，1998 Vol.12
[12]Shim J P,et al. Past, present, and future of decision support technology. DecisionSupport Systems ,June 2004 Vol.33 No.126
[13]陈莉，焦李成，Internet/Web数据挖掘研究现状及最新进展，西安电子科技大学学报（自然科学版），2001 Vol.1 No.
[14]林杰斌著，数据挖掘与OLAP理论与实务，北京：清华大学出版社
[15] R. Kosla and H. Blockeel,“Web mining research a survey,”SIG KDDExplorations, vol. 2, pp. 1–15, 2005 Vol. 5 No.28
[16] Michael J.A.Berry Gordon S.Linoff著,别荣芳,尹静,邓六爱译,数据挖掘技术,北京：机械工业出版社,2006.7
[17]Zhaohui tang Jamie MacLennan著,邝祝芳,焦贤龙等译,数据挖掘原理及应用,北京：清华大学出版社,2006.7
[18]尤玉林，张宪民，一种可靠的数据仓库中ETL策略与架构设计，计算机工程与应用，2005 Vol.21
[19]翁念龙，石晓晨，皮六一，多维建模数据管理方法及实现步骤，中国金融电脑，2005 Vol.8 No.9
[20] Guha Seal. CURE: An efficient clustering algorithm for large databases［A］. In:Proceedings of the ACM SIGMOD Conference on Management of Data: ACMPress. 1998 Vol.12 No.239.
[21]霍弗等著，现代数据库管理（第六版），电子工业出版社
[22]杨敏慧，张振华，用OLTP关系数据库管理系统分析数据，中国金融电脑，2004 Vol.5 No.40
[23]韩立平，庞瑞江，统计信息的革命科学决策的基石—记中国工商银行综合统计系统（CS2002），中国金融电脑，2005 Vol.2 No.13
[24]（意）朱迪茨著，袁方等译，实用数据挖掘，电子工业出版社
[25]李黛璐，分布式数据库技术在银行信贷综合管理系统中的应用，中国金融电脑，2005 Vol.5 No.9
[2]http://www.chonna.com/knowledge/news_article_display.asp?depid=2&subdepid=3&articleid=297淘金：数据挖掘在CRM中的运用
[3]王向星，袁胜，刘笑东，基于数据仓库的CRM在银行业中的应用，计算机世界.2003 Vol.28 No.16
[4]葛晨霞，商业银行CRM的需求分析和系统实现，中国金融电脑，2003 Vol.12No.9
[5]周意，李峰峰，基于CRM的银行业务应用模块分析，中国金融电脑，2004 Vol.9No.4
[6]高秩新，实现更具数据驱动力的银行客户关系管理，中国金融电脑，2005 Vol.6No.70
[7] Harjinder S.GILL等著,王仲谋,刘书丹译，数据仓库-客户服务器计算指南，北京：清华大学出版社,西蒙与舒斯特国际出版公司，1997 Vol.1
[8]陈文伟，黄金才，数据仓库与数据挖掘，北京：人民邮电出版社2004 Vol.1
[9]李子木,莫倩,周兴铭，数据库技术的研究现状及未来方向，计算机科学，1998Vol.25 No.4
[10] http://www.creawor.com/biforum/bi_03.htm，数据挖掘的研究现状
[11]王珊等，数据仓库技术与联机分析处理，科学出版社，1998 Vol.12
[12]Shim J P,et al. Past, present, and future of decision support technology. DecisionSupport Systems ,June 2004 Vol.33 No.126
[13]陈莉，焦李成，Internet/Web数据挖掘研究现状及最新进展，西安电子科技大学学报（自然科学版），2001 Vol.1 No.
[14]林杰斌著，数据挖掘与OLAP理论与实务，北京：清华大学出版社
[15] R. Kosla and H. Blockeel,“Web mining research a survey,”SIG KDDExplorations, vol. 2, pp. 1–15, 2005 Vol. 5 No.28
[16] Michael J.A.Berry Gordon S.Linoff著,别荣芳,尹静,邓六爱译,数据挖掘技术,北京：机械工业出版社,2006.7
[17]Zhaohui tang Jamie MacLennan著,邝祝芳,焦贤龙等译,数据挖掘原理及应用,北京：清华大学出版社,2006.7
[18]尤玉林，张宪民，一种可靠的数据仓库中ETL策略与架构设计，计算机工程与应用，2005 Vol.21
[19]翁念龙，石晓晨，皮六一，多维建模数据管理方法及实现步骤，中国金融电脑，2005 Vol.8 No.9
[20] Guha Seal. CURE: An efficient clustering algorithm for large databases［A］. In:Proceedings of the ACM SIGMOD Conference on Management of Data: ACMPress. 1998 Vol.12 No.239.
[21]霍弗等著，现代数据库管理（第六版），电子工业出版社
[22]杨敏慧，张振华，用OLTP关系数据库管理系统分析数据，中国金融电脑，2004 Vol.5 No.40
[23]韩立平，庞瑞江，统计信息的革命科学决策的基石—记中国工商银行综合统计系统（CS2002），中国金融电脑，2005 Vol.2 No.13
[24]（意）朱迪茨著，袁方等译，实用数据挖掘，电子工业出版社
[25]李黛璐，分布式数据库技术在银行信贷综合管理系统中的应用，中国金融电脑，2005 Vol.5 No.9