数据仓库技术在邮政金融客户系统中的研究与应用
|
摘要
随着国家宏观政策的调控，中国邮政分为邮政管理局和邮政企业集团，并且，进一步分为邮政银行和邮政企业集团。中国邮政银行必将面临更多的挑战与压力。如何有力的把握市场动态、有效的挽留金融客户、合理的部署渠道、准确的规避风险、合理的设置产品等将成为邮政储蓄将来更为紧迫的问题。另一方面，各业务系统积累了大量的客户资料、交易信息、渠道信息，这些信息充分反映了邮政储蓄业务发展的特点，隐含了邮政储蓄的经营现状和市场需求，使得邮政储蓄管理层可以从中发现有用的信息，用于进行企业的经营管理和分析、进行客户的差异化服务、进行针对性的营销活动，最大限度的提高企业的效益。正是在这种市场环境、业务要求和数据支撑的情况下，邮政储蓄提出了实施以数据仓库为基础的邮政金融客户信息管理系统。本文结合中国邮政金融客户管理的实际要求，开展了数据仓库技术在中国邮政金融客户系统的研究及应用。
本文首先介绍了对信息化发展所积累的海量数据进一步挖掘的意义和本课题的技术背景。然后根据金融客户系统的业务需求、储蓄业务系统的数据质量、数据仓库建设特点等因素，重点研究了金融客户系统的整体架构设计和模型设计两方面的内容，并给出了金融客户系统的系统架构设计、ETL设计、应用架构设计和逻辑模型及物理模型等。为了确保数据仓库的成功建设、保证数据仓库系统使用时的高效性，根据数据仓库的特点和性能上的要求，将逻辑数据模型转换成数据在物理设备上的存储结构与存取方法。
本文主要解决如下关键问题：从整体上给出了全国大集中的金融客户系统的系统整体架构，并在此基础上设计了系统逻辑体系架构、物理体系架构、应用逻辑体系等。基于整体架构设计重点研究设计了后台ETL数据处理机制，包括数据处理流程、数据质量检查和系统ETL的作业分类等。研究了前端应用总体设计，包含了应用的逻辑结构及对应的物理结构，同时还设计了系统的应该功能子系统。定制了符合系统实际情况的逻辑数据模型。考虑了数据库性能优化的有关处理方法。
本文介绍的系统架构、ETL处理机制、逻辑模型、物理模型、物理体系架构，已经在邮政金融客户管理系统(简称CPCIM)中投入使用。金融客户管理系统的成功上线，有力地推动了中国邮政金融系统管理的信息化，为经营决策提供强有力的数据支撑。
Along with our country's macro-control, State Postal Bureau of China divides into national postal industry and postal enterprises management, further more, divides into the china postal bank and the postal enterprises management. China postal bank will face to more challenge and pressure. As a result, there are more sever problems occurred: how to grasp marketplace development, persuade the finance customer to stay effectively, deploy a channel rationally, escape risk accurately, sets up a product rationally and so on. On the other hand, in every service system, there are plenty of customer data, transaction information, channel information, which reflect the characteristics of postal saving business, and imply the present managing status and the market need, so that the managing rank of postal saving business can find out some useful information and then apply them into the enterprises managing and different service for the customers, then, they can have the right marketing actions, and improve the enterprises profit in the large scale. Base on these situations, the postal saving business bring out the idea of carrying out Chinese postal service finance customer system which regards the data base as a foundation. Combined with the real demand, this article studies the application of the data storehouse technology in Chinese postal service finance customer system. The research job in this article will be helpful to the data storehouse technology application in Chinese postal service finance customer system and can directly guide the carrying out of Chinese postal service finance customer system.
This article firstly introduces the meaning of study the data accumulated in the informationization development and the technical background. Then, considering the business demand of the finance customer system, the data quality of the saving system and the data storehouse construction characteristic, this article mainly studies the overall frame design and model of finance customer system design, meanwhile, brings out the system frame, ELT designs, applies frame design and logic model and physics model etc. For insuring the success construction of the data storehouse and the high-efficiency of the data storehouse's application, based on the data storehouse's characteristics and properties, this article applies the logic data storehouse to the saving data in the physics equipment.
This article mainly solve the following key problems: First, gives a nationally frame of national gathering finance customer system, based on which designs the logic system frame, physics system frame, application logic system etc. Second, based on the nationally frame, designs the mechanism of the ETL background data handling, which includes data handling technological process, data mass examination and system ETL school assignment classification etc. Third, studies the fore-end application nationally designation that contains the applicative logic structure and corresponding physics structure, meanwhile designs the function subsystem. Fourth, designs the logic data model according with system reality. Last, takes the related processing methods of improving the data storehouse quality into account.
The system frame,, mechanism of the ETL background data processing, logic model , physics model , physics system frame , are already put in use in postal service finance customer management system (be called CPCIM for short). The success of Finance customer management system have forcefully driven the administrative informatization of Chinese postal service and provides the data for the managerial decision-making.
本文首先介绍了对信息化发展所积累的海量数据进一步挖掘的意义和本课题的技术背景。然后根据金融客户系统的业务需求、储蓄业务系统的数据质量、数据仓库建设特点等因素，重点研究了金融客户系统的整体架构设计和模型设计两方面的内容，并给出了金融客户系统的系统架构设计、ETL设计、应用架构设计和逻辑模型及物理模型等。为了确保数据仓库的成功建设、保证数据仓库系统使用时的高效性，根据数据仓库的特点和性能上的要求，将逻辑数据模型转换成数据在物理设备上的存储结构与存取方法。
本文主要解决如下关键问题：从整体上给出了全国大集中的金融客户系统的系统整体架构，并在此基础上设计了系统逻辑体系架构、物理体系架构、应用逻辑体系等。基于整体架构设计重点研究设计了后台ETL数据处理机制，包括数据处理流程、数据质量检查和系统ETL的作业分类等。研究了前端应用总体设计，包含了应用的逻辑结构及对应的物理结构，同时还设计了系统的应该功能子系统。定制了符合系统实际情况的逻辑数据模型。考虑了数据库性能优化的有关处理方法。
本文介绍的系统架构、ETL处理机制、逻辑模型、物理模型、物理体系架构，已经在邮政金融客户管理系统(简称CPCIM)中投入使用。金融客户管理系统的成功上线，有力地推动了中国邮政金融系统管理的信息化，为经营决策提供强有力的数据支撑。
Along with our country's macro-control, State Postal Bureau of China divides into national postal industry and postal enterprises management, further more, divides into the china postal bank and the postal enterprises management. China postal bank will face to more challenge and pressure. As a result, there are more sever problems occurred: how to grasp marketplace development, persuade the finance customer to stay effectively, deploy a channel rationally, escape risk accurately, sets up a product rationally and so on. On the other hand, in every service system, there are plenty of customer data, transaction information, channel information, which reflect the characteristics of postal saving business, and imply the present managing status and the market need, so that the managing rank of postal saving business can find out some useful information and then apply them into the enterprises managing and different service for the customers, then, they can have the right marketing actions, and improve the enterprises profit in the large scale. Base on these situations, the postal saving business bring out the idea of carrying out Chinese postal service finance customer system which regards the data base as a foundation. Combined with the real demand, this article studies the application of the data storehouse technology in Chinese postal service finance customer system. The research job in this article will be helpful to the data storehouse technology application in Chinese postal service finance customer system and can directly guide the carrying out of Chinese postal service finance customer system.
This article firstly introduces the meaning of study the data accumulated in the informationization development and the technical background. Then, considering the business demand of the finance customer system, the data quality of the saving system and the data storehouse construction characteristic, this article mainly studies the overall frame design and model of finance customer system design, meanwhile, brings out the system frame, ELT designs, applies frame design and logic model and physics model etc. For insuring the success construction of the data storehouse and the high-efficiency of the data storehouse's application, based on the data storehouse's characteristics and properties, this article applies the logic data storehouse to the saving data in the physics equipment.
This article mainly solve the following key problems: First, gives a nationally frame of national gathering finance customer system, based on which designs the logic system frame, physics system frame, application logic system etc. Second, based on the nationally frame, designs the mechanism of the ETL background data handling, which includes data handling technological process, data mass examination and system ETL school assignment classification etc. Third, studies the fore-end application nationally designation that contains the applicative logic structure and corresponding physics structure, meanwhile designs the function subsystem. Fourth, designs the logic data model according with system reality. Last, takes the related processing methods of improving the data storehouse quality into account.
The system frame,, mechanism of the ETL background data processing, logic model , physics model , physics system frame , are already put in use in postal service finance customer management system (be called CPCIM for short). The success of Finance customer management system have forcefully driven the administrative informatization of Chinese postal service and provides the data for the managerial decision-making.
引文
1．Hastie,T.,Tibshirani,R.,Friedman,J.著，范明，柴玉梅，昝红英等译．统计学习基础—数据挖掘、推理与预测．北京：电子工业出版社，2004．1．
2．Tom M.Mitchell 著，曾华军，张银奎等译．机器学习．北京：机械工业出版社，2003．1
3．J.Han and M.Karnber 著，范明，孟小峰译．数据挖掘：概念与技术．北京：机械工业出版社．2001
4．企业级数据仓库原理、设计与实践，王彦龙，电子工业出版社，2006
5．数据挖掘：概念与技术 作者：(加)[韩家炜]JiaweiHan，(加)Micheline Kamber著；范明，孟小峰等译
6．数据仓库基础，Paulraj Ponniah，电子工业出版社，2004
7．数据仓库项目管理实践，于宗民等，人民邮电出版社
8. Using the Data Warehouse. Inmon, W.H. and Richard Hackathorn.,New York: John Wiley & Sons. 1994
9．OLAP解决方案：创建多维信息系统，Erik Thomsen，电子工业出版社，2004；第二版
10. The Data Warehouse Toolkit, Kimball, R., New York: John Wiley & Sons.1996
11. Designing a Data Warehouse Supporting Customer Relationship Management, Chris Todman, Pearson Education Press
12. Building the Data Warehouse, W.H. Inmon, New York: John Wiley & Sons. 1996
13. Document Warehousing and Text Mining. Sullivan, Dan., New York: John Wiley & Sons. 2001
14. Frean M. The upstar algorithm: A method for constructing and training feed-forward neural networks Neural Computation , 1990.
15. Yao X. A review of evolutionary artificial neural networks International Journal of Intelligent Systems , 1993.
16. Maniezzo V. Genetic evolution of the topology and weight distribution of neural networks IEEE Trans On Neural Networks, 1994.
17. W H Inmon. building the Data Warehouse New York:John Wiley &Sons, 1993.
18. Mattison, Nob. Warehousing Wherewithal CIO , 1996.
19. Inmon W H. The Data Warehouse and Data Mining Communication of the ACM, 1996
20．数据仓库技术及其应用[科技创业月刊Pioneering with Science ＆ Technology Monthly]晏伯武
21．数据挖掘技术在CRM中应用浅析[J]张红岩 现代情报，2005，(12)
22．数据挖掘工具的评判[J]．刘世平，姚玉辉．数字财富，2003，(06)
23．国外先进数据挖掘工具的比较分析[J]张雪英计算机工程，2003，(16)
24．建立客户关系管理迫在眉睫[J]．孙晶 出版广角，2002，(09)
25．数据挖掘之道[J]．李孔仁，王成才，王登科．牡丹江师范学院学报(自然科学版)，2005，(01)
26．数据挖掘技术在商业银行客户关系管理中的应用分析[J]．黄华卿，张维，熊熊．哈尔滨商业大学学报(社会科学版)，2006，(03)
27．数据挖掘工具的评判[J] 周健吉 电脑知识与技术，2004，(23)
28．浅谈我国银行客户关系管理系统的建设[J] 张幼玲，余殿芳 华南金融电脑，2005，(11)
29．数据挖掘及其在电信客户关系管理中的应用[J] 李建福 重庆科技学院学报(自然科学版)，2006，(01)
30．CRM：新经济时代的经营策略[J] 魏中龙 北京商学院学报，2001，(03)
31．中小企业实现CRM的难点及对策[J] 闫丽霞 商场现代化，2005，(24)
32．客户关系管理中的数据挖掘[J] 吴志军 商场现代化，2005，(20)
33．CRM实施的几个问题探讨[J] 刘伟 信息技术与信息化，2005，(06)
34．数据挖掘技术在客户关系管理中的应用研究[J] 戴艳红，贺红燕．商场现代化，2006，(34)
35．浅谈如何建立有效的客户关系管理[J] 曹锦川 甘肃科技纵横，2005．(04)
36．如何在我国商业银行中有效实施 CRM 的研究[J]．程社欣，程春明．内蒙古石油化工，2006，(06)
37．我国银行导入CRM的战略思考[J]．盛丽．武汉金融，2002，(04)
38．CRM系统在企业网络营销中的作用[J]．潘伟．中国流通经济，2001，(02)
39．关于我国企业目前实施 CRM 的理性思考[J]．宋思根．中央财经大学学报，2002．(11)
40. S.Russell, J.Binder, D. Koller, and K.Kanazawa. Local learning in probabilistic networks with hidden variables. In Proc. 14th Joint lnt. Conf. On Artificial Intelligence (IJCAI'95), volume 2, Pages: 1146-1152, Montreal, Canada, Aug. 1995.
41. J.R.Quinlan. Simplifying decision trees. Internation Journal of Man-Machine Studies, 27.Pages: 221-234,1987.
42. M.Mehta, R. Agrawal, and J.Rissanen. SLIQ: A fast scalable classifier for data mining. In Proc. 1996 Int. Conf. Extendding Database Technology (EDBT'96), Avignon, France, Mar. 1996.
43. J.C.Shafer, R. Agrawal, M.Mehta: SPRINT: A Scalable classifier Parallel Classifier for Data Mining.In Proceedings of the 22th international Conference on Very Large Databases, Mumbai (Bombay), India, Sept. 1996.
44. P.Domingos and M.Pazzani. Beyond independence: Conditions for the optimality of the simple Bayesian classifier. In Proc. 13th Intl. Conf. Machine Learning, Pages 105-112, 1996.
45. B.Liu, W.Hsu, and Y.Ma. Integrating classification and association rule mining. In Proc.1998 Int. Conf. Knowledge Discovery and Data Mining (KDD'98), Pages 80-86, New York, Aug. 1998.
46. Li W, Han J, and Pei J. CMAR: Accurate and efficient classification based on multiple class-association rules. In ICDM'01, San Jose, CA, Nov. 2001, Pages. 369-376
47. Xiaoxin Yin, Jiawei Han: CPAR: Classification based on Predictive Association Rules. SDM 2003
48. Jinyan Li, Guozhu Dong, Kotagiri Rarnamohanarao, Limsoon Wong: DeEPs: A New Instance-Based Lazy Discovery and Classification System. Machine Learning 54(2): Pages 99-124,2004.
2．Tom M.Mitchell 著，曾华军，张银奎等译．机器学习．北京：机械工业出版社，2003．1
3．J.Han and M.Karnber 著，范明，孟小峰译．数据挖掘：概念与技术．北京：机械工业出版社．2001
4．企业级数据仓库原理、设计与实践，王彦龙，电子工业出版社，2006
5．数据挖掘：概念与技术 作者：(加)[韩家炜]JiaweiHan，(加)Micheline Kamber著；范明，孟小峰等译
6．数据仓库基础，Paulraj Ponniah，电子工业出版社，2004
7．数据仓库项目管理实践，于宗民等，人民邮电出版社
8. Using the Data Warehouse. Inmon, W.H. and Richard Hackathorn.,New York: John Wiley & Sons. 1994
9．OLAP解决方案：创建多维信息系统，Erik Thomsen，电子工业出版社，2004；第二版
10. The Data Warehouse Toolkit, Kimball, R., New York: John Wiley & Sons.1996
11. Designing a Data Warehouse Supporting Customer Relationship Management, Chris Todman, Pearson Education Press
12. Building the Data Warehouse, W.H. Inmon, New York: John Wiley & Sons. 1996
13. Document Warehousing and Text Mining. Sullivan, Dan., New York: John Wiley & Sons. 2001
14. Frean M. The upstar algorithm: A method for constructing and training feed-forward neural networks Neural Computation , 1990.
15. Yao X. A review of evolutionary artificial neural networks International Journal of Intelligent Systems , 1993.
16. Maniezzo V. Genetic evolution of the topology and weight distribution of neural networks IEEE Trans On Neural Networks, 1994.
17. W H Inmon. building the Data Warehouse New York:John Wiley &Sons, 1993.
18. Mattison, Nob. Warehousing Wherewithal CIO , 1996.
19. Inmon W H. The Data Warehouse and Data Mining Communication of the ACM, 1996
20．数据仓库技术及其应用[科技创业月刊Pioneering with Science ＆ Technology Monthly]晏伯武
21．数据挖掘技术在CRM中应用浅析[J]张红岩 现代情报，2005，(12)
22．数据挖掘工具的评判[J]．刘世平，姚玉辉．数字财富，2003，(06)
23．国外先进数据挖掘工具的比较分析[J]张雪英计算机工程，2003，(16)
24．建立客户关系管理迫在眉睫[J]．孙晶 出版广角，2002，(09)
25．数据挖掘之道[J]．李孔仁，王成才，王登科．牡丹江师范学院学报(自然科学版)，2005，(01)
26．数据挖掘技术在商业银行客户关系管理中的应用分析[J]．黄华卿，张维，熊熊．哈尔滨商业大学学报(社会科学版)，2006，(03)
27．数据挖掘工具的评判[J] 周健吉 电脑知识与技术，2004，(23)
28．浅谈我国银行客户关系管理系统的建设[J] 张幼玲，余殿芳 华南金融电脑，2005，(11)
29．数据挖掘及其在电信客户关系管理中的应用[J] 李建福 重庆科技学院学报(自然科学版)，2006，(01)
30．CRM：新经济时代的经营策略[J] 魏中龙 北京商学院学报，2001，(03)
31．中小企业实现CRM的难点及对策[J] 闫丽霞 商场现代化，2005，(24)
32．客户关系管理中的数据挖掘[J] 吴志军 商场现代化，2005，(20)
33．CRM实施的几个问题探讨[J] 刘伟 信息技术与信息化，2005，(06)
34．数据挖掘技术在客户关系管理中的应用研究[J] 戴艳红，贺红燕．商场现代化，2006，(34)
35．浅谈如何建立有效的客户关系管理[J] 曹锦川 甘肃科技纵横，2005．(04)
36．如何在我国商业银行中有效实施 CRM 的研究[J]．程社欣，程春明．内蒙古石油化工，2006，(06)
37．我国银行导入CRM的战略思考[J]．盛丽．武汉金融，2002，(04)
38．CRM系统在企业网络营销中的作用[J]．潘伟．中国流通经济，2001，(02)
39．关于我国企业目前实施 CRM 的理性思考[J]．宋思根．中央财经大学学报，2002．(11)
40. S.Russell, J.Binder, D. Koller, and K.Kanazawa. Local learning in probabilistic networks with hidden variables. In Proc. 14th Joint lnt. Conf. On Artificial Intelligence (IJCAI'95), volume 2, Pages: 1146-1152, Montreal, Canada, Aug. 1995.
41. J.R.Quinlan. Simplifying decision trees. Internation Journal of Man-Machine Studies, 27.Pages: 221-234,1987.
42. M.Mehta, R. Agrawal, and J.Rissanen. SLIQ: A fast scalable classifier for data mining. In Proc. 1996 Int. Conf. Extendding Database Technology (EDBT'96), Avignon, France, Mar. 1996.
43. J.C.Shafer, R. Agrawal, M.Mehta: SPRINT: A Scalable classifier Parallel Classifier for Data Mining.In Proceedings of the 22th international Conference on Very Large Databases, Mumbai (Bombay), India, Sept. 1996.
44. P.Domingos and M.Pazzani. Beyond independence: Conditions for the optimality of the simple Bayesian classifier. In Proc. 13th Intl. Conf. Machine Learning, Pages 105-112, 1996.
45. B.Liu, W.Hsu, and Y.Ma. Integrating classification and association rule mining. In Proc.1998 Int. Conf. Knowledge Discovery and Data Mining (KDD'98), Pages 80-86, New York, Aug. 1998.
46. Li W, Han J, and Pei J. CMAR: Accurate and efficient classification based on multiple class-association rules. In ICDM'01, San Jose, CA, Nov. 2001, Pages. 369-376
47. Xiaoxin Yin, Jiawei Han: CPAR: Classification based on Predictive Association Rules. SDM 2003
48. Jinyan Li, Guozhu Dong, Kotagiri Rarnamohanarao, Limsoon Wong: DeEPs: A New Instance-Based Lazy Discovery and Classification System. Machine Learning 54(2): Pages 99-124,2004.