【作者】 楚荣珍；

【导师】 张新荣； 宋栓虎；

【作者基本信息】 天津大学 ， 计算机技术， 2007， 硕士

【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。更多还原

【Abstract】 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.更多还原