用户名: 密码: 验证码:
银行卡客户细分系统分析与设计
详细信息    本馆镜像全文|  推荐本文 |  |   获取CNKI官网全文
摘要
随着信贷等传统业务的利润空间逐渐缩小,现代商业银行更加注重发展中间业务。其中,银行卡业务作为一项重要的中间业务,已成为商业银行的竞争焦点。目前在西方发达国家,银行卡业务是许多国际大银行的主要业务和主要利润来源。我国商业银行的银行卡业务起步较晚,由于受各方面条件的制约,在银行卡业务的经营方式、风险控制、产品种类及服务等各方面都与国外发卡银行有较大的差距。
    目前国内银行卡市场的竞争日益激烈。一方面,加入WTO后,外资银行进军我国,由于在营业网点上并不具有优势,它们将银行卡业务作为与国内银行业竞争的首要目标。另一方面,中国银联成立后,我国银行卡实现了全国联网通用,各中小商业银行也抓住此契机大力发展银行卡业务。
    我国银行卡产业目前的主要问题是发卡银行的经营理念落后,长期以来一直将银行卡定位为电子化银行存折,发卡的主要目的是为信贷等业务筹措资金。在这种思想的指导下,国内各发卡银行只是一味的追求银行卡发行量,盲目发卡,而未对客户群体进行细分,为客户提供有针对性的银行卡产品和服务。这种做法一方面造成了大量持卡人用卡积极性不高,已发行的银行卡中很大一部分成为死卡、睡眠卡,造成了发卡银行不必要的资源浪费。另一方面,未对客户进行细分也使得国内发卡银行信用卡业务的风险居高不下,极大的限制了国内发卡银行信用卡业务的发展,导致银行卡业务的利润水平不高。返观国外各大发卡银行,它们在经营银行卡业务过程中非常注重对其客户数据进行分析利用。随着统计学、人工智能及计算机等各方面的技术的发展,越来越多的发卡银行开始采用数据挖掘这一新技术对其海量业务数据进行挖掘,从中发掘有用的商业知识,以辅助其决策。它们通过客户细分区分客户群体,在此基础
    
    
    上有针对性的设计与发行银行卡产品,实施客户关系管理,同时有针对性的进行客户风险控制。
    因此,我国发卡银行要提高自身的市场竞争力,必须转变以往的经营理念,借鉴国外发卡银行的经验,实施客户细分,划分客户群体。目前国内部分发卡银行借鉴国外发卡银行的已有做法,针对特定的女性客户群体和校园客户群体发行了专用银行卡,在银行卡客户细分方面迈出了第一步。但是国内发卡银行仍然普遍缺少一套基于数据驱动的、科学客观的客户细分工具。
    鉴于我国商业银行大量业务数据存放在数据库中未能得到有效的分析和利用的情况,中国人民银行科技司组织了国家“十五”科技攻关课题“电子银行模拟系统”。该课题的主要目的是利用现代信息技术、统计学原理和人工智能技术,建立科学的、客观的和先进的电子银行模拟客户分析与风险评价系统,逐步推广应用到国内商业银行,为商业银行客户细分、客户价值发现、客户服务、市场策划、客户风险管理等提供科学的、客观的决策支持。该课题由中国人民银行模拟银行实验中心承办,已于2003年10月份通过了课题验收组的最终验收。
    笔者有幸参加了“电子银行模拟系统”课题的研究及系统开发工作,主要负责其中的银行卡客户细分子系统的分析与设计工作。在课题研究过程中,笔者前往中国农业银行四川省分行进行了调研工作,从农行相关部门获得了实际的客户数据,然后运用SAS软件中的数据挖掘工具Enterprise Miner,对农行的客户数据进行了数据挖掘分析,在此基础上设计并开发了银行卡客户细分系统。这便是本论文的写作背景。
    本论文总结了我国银行卡产业的发展历程,分析了我国银行卡产业目前存在的问题,从而提出了国内发卡银行实施客户细分的必要性,并分析了国内外发卡银行实施客户细分的情况。在阐述数据挖掘的主要概念、解决的问题及使用的方法、过程及工具等的基础上,介绍了笔者如何运用SAS的数据挖掘工具Enterprise Miner对农业银行四川省分行的银行卡客户数据进行数据挖掘建模分析。最后
    
    
    阐述了在数据挖掘的基础上对银行卡客户细分系统进行分析和设计的过程。
    本论文主要分为四章。第一章主要分析了国内发卡银行实施客户细分的必要性。本章首先介绍了银行卡的基本概念及分类,回顾了我国银行卡产业的发展历程,分析了我国银行卡市场的现状及存在的主要问题,从而提出了银行卡客户细分的必要性;接着对客户细分的理论进行了简单介绍,并分析了银行卡客户细分的意义及细分的标准;最后总结了国内外发卡银行在实施客户细分方面的情况。
    第二章是本论文的核心部分,主要是阐述笔者运用数据挖掘技术对中国农业银行四川省分行的银行卡客户数据进行数据挖掘分析的过程。本章首先简单介绍了数据挖掘的基本概念,分别阐述了数据挖掘解决的各类问题及使用的各种方法。在综合对比分析几种流行的数据挖掘方法论的基础上归纳总结了数据挖掘分析的一般过程。然后介绍了笔者如何运用Enterprise Miner对农行银行卡客户数据进行数据挖掘分析,其中,对于银行卡客户预测性细分采用了决策树方法,对银行卡客户聚类采用了Kohonen自组织特征映射神经网络方法。
    第三章是在第二章数据挖掘分析结果的基础上,阐述银行卡客户细分系统的设计过程。本章首先介绍了该系统的开发背景、主要功能及开发环境的选择,然后按照软件工程的基本思路依次?
With the gradual profit reducing of traditional businesses, commercial banks pay more and more attention to develop intermediary business. Nowadays, the banking card business is the most important business and the main profit source of commercial banks in western countries. Chinese commercial banks’ banking card business has been started late. Being restricted by various factors, there is still large gap between domestic commercial banks and foreign ones.
    The competition of domestic banking card market is becoming fiercer and fiercer. On one hand, after China’s entry into the WTO, the foreign commercial banks are expanding their business in Chinese by developing their banking card business at first. On the other hand, after the establishment of Chinaunionpay (CUP) Inc., the CUP banking cards can be used through various commercial banks in China. Capturing this opportunity, many small commercial banks in China are beginning to develop their banking card businesses.
    The main problem of local banking card industry is that these issuing banks’ ideas are behind the time. They always regard banking card as an electronic from deposit book, which raise money for the credit business. Being guided by this thought, these local issuing banks just pursued the growth of the number of issued banking card, without segment their customers. As a result, the cardholders are not positive to use their banking cards, which don’t meet their demands well. Additionally, lacking of customer segmentation, the credit card business of local issuing banks is running high risks.
    In developed countries, most of these issuing banks put great emphasis on the analysis of customer data. With the development of statistics, artificial intelligence technology and information technology, more and more issuing banks utilize the data mining technology to find
    
    
    valuable knowledge in their huge amount of data, and use the results to supply their decisions. By segmenting their customers, they design proper banking card and carry out customer relationship management. Besides, they can control the risk more effectively.
    Therefore, to enhance their competition, domestic issuing banks must change their ideas and segment their customers. So far some domestic issuing banks have taken some customer segmentation strategies, making use of experience of foreign issuers. However, domestic issuing banks still lack of a set of tools that segment customers objectively and scientifically.
    The science and technology bureau of the people's bank of China organized the research work of the "electronic bank simulation system". The purpose of this research is to develop scientifically, objectively computerized customer analysis and risk-valuing system by utilizing information technology, statistics and artificial intelligence technology. This research was undertaken by the experimental center of the People's Bank of China and has passed the final examination and acceptance in October 2003. The author of this thesis was fortunate to participate in the research work and was assigned to the work of analyzing and designing the banking customer segmentation subsystem. During the research, the research team went to the Agricultural Bank of China (ABC) Sichuan branch for survey. The author used the data mining tool of SAS—Enterprise Miner to mine the practical customer data from ABC Sichuan branch. Based on the data mining process, the author designed and developed the banking card customer system. All of these is the background of this thesis.
    This thesis is composed of four chapters. In the first chapter, the author analyzed the necessity of customer segmentation of issuing banks. Firstly, it introduced the basic concept and classifies of banking card, then reviewed the development history of Chinese banking card industry.
    
    
    After that, the theory of customer segmentation was introduced simply, together with the significance and standard of banking card customer segmentation analyzed. Finally the author summarized the customer segmentation of domestic and
引文
[加]Jiawei Han, Micheline Kamber 著,范明,孟小峰 等译. 数据挖掘:概念与技术,机械工业出版社,2001
    [美]Alex Berson,Stephen Smith,Kurt Thearling 著,贺奇,郑岩,魏黎,蔡致远等 译. 《构建面向CRM的数据挖掘应用》,人民邮电出版社,2001
    王学民. 《应用多元统计》(第二版),上海财经大学出版社,2004
    苏建文. 银行卡差异化营销策略,《中国信用卡》,2002.1
    李黄燕. 以客户为中心的银行卡经营理念,《中国信用卡》,2001.10
    赵勇. 银行卡产品的市场细分,《中国信用卡》,2001.11
    李建平. 信用卡客户数据库营销,《中国信用卡》,2000.2
    张阿兰,谢邦昌. 数据挖掘在信用卡公司的应用,《统计与信息论坛》,2002.11
    王越,曹长修. DM技术在信用卡管理中的应用,《计算机工程与应用》,2002.10
    唐华松,姚耀文. 数据挖掘中决策树算法的探讨,《计算机应用研究》,2001.8
    薛薇. 数据挖掘中的决策树技术及其应用,《统计与信息论坛》 2002.3
    行小帅,焦李成. 数据挖掘的聚类方法,《电路与系统学报》,2003.2
    黄亦潇,李箐箐,邵培基. 基于SOM人工神经网络的客户资源分类方法研究,《重庆大学学报》,2003.6
    Enterprise Miner Tutorial. SAS Inc.
    Mesbah Ahmed, Chris Garrett, Jerymy Faircloth, Chris Payne, Wei Meng Lee, Jonothon Ortiz. ASP.NET Web Developer's Guide,
    
    
    Syngress Publishing Inc.,2002
    刘信彦. 《ASP.NET技术参考》,中国铁道出版社,2001
    朱斌,宋先忠. 动态网页开发技术探讨,《计算机应用》,2001.9
    薛华成.《管理信息系统》(第三版),清华大学出版社,1999
    萨师煊,王珊.《数据库系统概论》(第二版),高等教育出版社,1991

© 2004-2018 中国地质图书馆版权所有 京ICP备05064691号 京公网安备11010802017129号

地址:北京市海淀区学院路29号 邮编:100083

电话:办公室:(+86 10)66554848;文献借阅、咨询服务、科技查新:66554700