数据挖掘在通信行业CRM中的应用研究
|
摘要
一个企业要生存，客户是最重要的资源。由于买方市场格局进一步强化，来自市场的压力不断加大，争取客户忠诚度的竞争进一步加剧，企业不得不更深入地考虑如何争取和保持客户，以显著提高其销售能力。针对这种情况，CRM(客户关系管理)应运而生。一个成功的CRM在很大程度上可以解决企业的对于客户的服务满意度，但从另一个方向考虑，我们可以让CRM为企业自身做得更多，通过利用数据挖掘技术分析CRM累积的数据，可以让企业更好的利用其客户资源。
数据仓库的主要功能是提供企业决策支持系统(DSS)或行政信息系统(EIS)所需要的信息，它把企业日常营运中分散不一致的数据经归纳整理之后转换为集中统一的、可随时取用的深层信息。同时，由于数据与信息量成倍的增长，人们迫切需要一种有效的工具来利用这些数据，从中提取所需的知识和作进一步分析。这实际上是一知识发现的过程。数据挖掘就是这一过程的核心技术。数据挖掘是利用已有数据的一种新型、有效的技术。它已引起了研究领域、工业领域和媒体广泛关注。数据挖掘(Data Mining)就是从大量的、不完全的、有噪声的、模糊的、随机的数据中，提取隐含在其中的、人们事先不知道的、但又是潜在有用的信息和知识的过程。原始数据可以是结构化的，如关系数据库中的数据，也可以是半结构化的，如文本、图形、图像数据，甚至是分布在网络上的异构型数据。发现知识的方法可以是数学的，也可以是非数学的；可以是演绎的，也可以是归纳的。发现了的知识可以被用于信息管理、查询优化、决策支持、过程控制等，还可以用于数据自身的维护。
本文主要研究了通信行业中利用数据挖掘技术对CRM累积的数据进行分析，从而获得本企业和竞争对手的相关竞争能力和趋势的结果，实际应用中证明分析结果是可信的、有价值的。全文共分为五章，第一章前言，主要介绍近年来国内外对CRM、数据挖掘等方面的研究现状。第二章挖掘CRM数据，介绍了数据仓库和数据挖掘的概念，分析了原系统中数据挖掘的不足之处，并提出几种常用的数据挖掘方法，根据通信行业的数据特点进行了选择。第三章关联规则介绍了在选择了数据挖掘方法后使用的算法以及在其过程中涉及的问
通信行业CRM中数据挖掘的应用研究
题。第四章列举了数据分析的结果。从结果看数据挖掘是成功的，其
结果是可信的。第五章总结了数据挖掘应用于cRM数据的成果和广
泛应用前景。
Generally speaking, more or less, because of signal and substance transferring, time-delay problem does exist in any system. With regard to continuous time-delay system, it's infinitude dimensions in theory. As for discrete time-delay system, the dimensions of problem will increase progressionally with time-delay. Therefore it's difficult both mathematically and practically for us to analyze and synthesize the system with time-delay. As a matter of fact, in order to avoid the difficulties of analyzing time-delay system, we often ignore the effects of small time-delay without sacrificing the system precision. But when it comes to some systems which can not be ignored, we must take the effects of time-delay into consideration when analyzing and synthesizing the systems. In addition to those systems, for the sake of avoiding the difficulties of analyzing higher order systems, we often simulate the system with one order time-delay model. So we can draw a conclusion the analysis and synthesis of time-delay system
is of great values both in academics and practice, also it's one of the hardest problems in control system.
For discrete time-delay system, a optimal control problem on quadratic performance indexes are often transformed into a two-point boundary value problem with not only time-delay terms but time-advance terms. It's difficult to solve this optimal control problem for exact solutions or numeral solutions. Another way of solving optimal problem of discrete time-delay is treating all time-delay terms as status variable, then transform the system into higher order non-delay system. The cost of this method is that with the prolong of time-delay, it may cause vicious increasing of system orders, especially for long time-delay system. Solving the suboptimal problem of discrete time-delay system is of great significance.
In this thesis, the suboptimal control of linear discrete time-delay system is mainly studied. In chapter 1, we introduce the study status in quo(the stability, the stabilization, the robustness, the guaranteed cost
control, the forecast control, the optimal control and the suboptimal control problem on time-delay systems). In chapter 2 the methods we have often seen are introduced in solving the suboptimal control problem. In chapter 3, for suboptimal control problems of linear discrete time-delay system, a new method of suboptimal control based on non-delay transformation approach is proposed. By this method, we regard time-delay terms as additional disturbance for two-point boundary problem with time-delay terms and time-advance terms, and introduce interactive approach arithmetic to the problem. By this means, we can transform a two-point boundary problem with time-delay terms and time-advance terms into a repeated approach problem without time-delay terms and time-advance terms. When we regard the Nth iterative resolution as optimal law of the system, we obtain suboptimal law. Simulation result shows that the algorithmic is effectual. In chapter 4, we apply the non-delay transform method to a linear discrete large-
scale system with small time-delay. First, we decompose large-scale system into a group of decoupled subsystem with additional distribution. Second, we transform higher order two-point boundary problem with time-delay terms and time-advance terms into a group of decoupled problem without time-delay terms and time-advance terms. By taking the Mth interactive values of optimal law as that of the suboptimal control of large-scale system. For linear large-scale systems with small time-delay, interactive times of computing suboptimal control law may be greatly decreased. Therefore, this approach is also specially suited for the suboptimal controller design for the linear large-scale systems with small time-delay.
数据仓库的主要功能是提供企业决策支持系统(DSS)或行政信息系统(EIS)所需要的信息，它把企业日常营运中分散不一致的数据经归纳整理之后转换为集中统一的、可随时取用的深层信息。同时，由于数据与信息量成倍的增长，人们迫切需要一种有效的工具来利用这些数据，从中提取所需的知识和作进一步分析。这实际上是一知识发现的过程。数据挖掘就是这一过程的核心技术。数据挖掘是利用已有数据的一种新型、有效的技术。它已引起了研究领域、工业领域和媒体广泛关注。数据挖掘(Data Mining)就是从大量的、不完全的、有噪声的、模糊的、随机的数据中，提取隐含在其中的、人们事先不知道的、但又是潜在有用的信息和知识的过程。原始数据可以是结构化的，如关系数据库中的数据，也可以是半结构化的，如文本、图形、图像数据，甚至是分布在网络上的异构型数据。发现知识的方法可以是数学的，也可以是非数学的；可以是演绎的，也可以是归纳的。发现了的知识可以被用于信息管理、查询优化、决策支持、过程控制等，还可以用于数据自身的维护。
本文主要研究了通信行业中利用数据挖掘技术对CRM累积的数据进行分析，从而获得本企业和竞争对手的相关竞争能力和趋势的结果，实际应用中证明分析结果是可信的、有价值的。全文共分为五章，第一章前言，主要介绍近年来国内外对CRM、数据挖掘等方面的研究现状。第二章挖掘CRM数据，介绍了数据仓库和数据挖掘的概念，分析了原系统中数据挖掘的不足之处，并提出几种常用的数据挖掘方法，根据通信行业的数据特点进行了选择。第三章关联规则介绍了在选择了数据挖掘方法后使用的算法以及在其过程中涉及的问
通信行业CRM中数据挖掘的应用研究
题。第四章列举了数据分析的结果。从结果看数据挖掘是成功的，其
结果是可信的。第五章总结了数据挖掘应用于cRM数据的成果和广
泛应用前景。
Generally speaking, more or less, because of signal and substance transferring, time-delay problem does exist in any system. With regard to continuous time-delay system, it's infinitude dimensions in theory. As for discrete time-delay system, the dimensions of problem will increase progressionally with time-delay. Therefore it's difficult both mathematically and practically for us to analyze and synthesize the system with time-delay. As a matter of fact, in order to avoid the difficulties of analyzing time-delay system, we often ignore the effects of small time-delay without sacrificing the system precision. But when it comes to some systems which can not be ignored, we must take the effects of time-delay into consideration when analyzing and synthesizing the systems. In addition to those systems, for the sake of avoiding the difficulties of analyzing higher order systems, we often simulate the system with one order time-delay model. So we can draw a conclusion the analysis and synthesis of time-delay system
is of great values both in academics and practice, also it's one of the hardest problems in control system.
For discrete time-delay system, a optimal control problem on quadratic performance indexes are often transformed into a two-point boundary value problem with not only time-delay terms but time-advance terms. It's difficult to solve this optimal control problem for exact solutions or numeral solutions. Another way of solving optimal problem of discrete time-delay is treating all time-delay terms as status variable, then transform the system into higher order non-delay system. The cost of this method is that with the prolong of time-delay, it may cause vicious increasing of system orders, especially for long time-delay system. Solving the suboptimal problem of discrete time-delay system is of great significance.
In this thesis, the suboptimal control of linear discrete time-delay system is mainly studied. In chapter 1, we introduce the study status in quo(the stability, the stabilization, the robustness, the guaranteed cost
control, the forecast control, the optimal control and the suboptimal control problem on time-delay systems). In chapter 2 the methods we have often seen are introduced in solving the suboptimal control problem. In chapter 3, for suboptimal control problems of linear discrete time-delay system, a new method of suboptimal control based on non-delay transformation approach is proposed. By this method, we regard time-delay terms as additional disturbance for two-point boundary problem with time-delay terms and time-advance terms, and introduce interactive approach arithmetic to the problem. By this means, we can transform a two-point boundary problem with time-delay terms and time-advance terms into a repeated approach problem without time-delay terms and time-advance terms. When we regard the Nth iterative resolution as optimal law of the system, we obtain suboptimal law. Simulation result shows that the algorithmic is effectual. In chapter 4, we apply the non-delay transform method to a linear discrete large-
scale system with small time-delay. First, we decompose large-scale system into a group of decoupled subsystem with additional distribution. Second, we transform higher order two-point boundary problem with time-delay terms and time-advance terms into a group of decoupled problem without time-delay terms and time-advance terms. By taking the Mth interactive values of optimal law as that of the suboptimal control of large-scale system. For linear large-scale systems with small time-delay, interactive times of computing suboptimal control law may be greatly decreased. Therefore, this approach is also specially suited for the suboptimal controller design for the linear large-scale systems with small time-delay.
引文
[1]R. Agrawal, T. Imielinski, and A. Swami. Mining association rules between sets of items in large databases. Proceedings of the ACM SIGMOD Conference on Management of data, pp. 207-216, 1993.
[2]李东主编，企业信息化案例，北京大学出版社，P16-34
[3]Charles E．Spurgenon著，张健／袁晓辉／苗沛荣译，以太网技术入门与实现，机械工业出版社P1-2
[4]孙伟纲，企业战略管理与管理驾驶舱，计算机世界，1999-11-15 C19
[5]Hohhof. B. Developing, Information system for Competitiveintelli-gence, Librarytrends. Vol. 43, No. 2, Fall, 1994, P226-238
[6]刘兴，敢问路在何方(对我国企业MIS建设的思考)，IT经理世界，1999年第12期
[7]吕春明，信息化：决策精确之路，中国计算机报，2001-09-06，P67
[8][美]艾森．拉塞尔／保罗．弗里嘉著 张涛／赵陵 译，麦肯锡意识，华夏出版社，P64
[9]Michaele Porter，陈小悦译，竞争战略，华夏出版社
[10]Michaele Porter著，陈小悦译，竞争优势，华夏出版社
[11]尼葛洛庞帝著，胡泳／范海燕译，数字化生存，海南出版社
[12]陈志宏，CRM：顾客关系管理，经济管理，2001年，第9期
[13]丁宁，信息技术革命与企业组织创新，经济管理出版社，2001年
[14]刘首英主编，Dialogue：Listening to the Voice of World's Top Business Leaders，中国发展出版社
[15][美]David Hand等著 张银奎等译，数据挖掘原理，机械工业出版社，2003
[16][美]W．H．Inmon著 王志海 林友芳等译，Building the Data Warehouse，机械工业出版社，2003
[17][美]Alex Berson Ste，构建面向CRM的数据挖掘应用，清华大学出版社，2001
[18]Michael F.S. Chan/Walter W.C. Chung, A Framework to Develop an Enterprise Information Portal for Contract Manufacturing, Int. J. Production Economics 75(2002) 113-126
[19]F. Bezzo/S. Macchietto/C. C. Pantelides, A General Framework for the Integration of Computational Fluid Dynamics and Process Simulation Computers and Chemical Engineerine. 24 (2000)
[20]D.H.R. Price/R. Beach/A.P. Muhlemann/J.A. Sharp/A. Paterson, A System to Support the Enhancement of Strategic Qexibility in Manufacturing Enterprises, European Journal of Operational Research 109(1998) 362-376
[21]Sung HoHa/Sang Chun Park, Application of Data Mining Tools to Hotel Data Mart on the Intranet for Database Marketing, Experts System With Application 15(1998) 1-3
[22]M. Levya/P. Powell/R. Galliers, Assessing Information Systems Strategy Development Frameworks in SMEs, Information & Management 35 (1999) 247-251
[23]R. Agrawal, and J. Shafer. Parallel mining of association rules:Design,Implementaion, and Experience. Technical Repoort FJ10004, IBM
Almaden Research Center, San Jose, CA 95120, Jan. 1996.
[24]S. Brin, R. Motwani, and C. Silverstein. Beyond market baskets:generlizing association rules to correlations. Proceedings of the ACM SIGMOD, 1996. pages 255-276.
[25]S. Brin, R. Motwani, J. D. Ullman, and S. Tsur. Dynamic Itemset counting and implication rules for market basket data. In ACM SIGMOD International Conference On the Management of Data. 1997.
[26]F. Korn, A. Labrinidis, Y. Kotidis, and C. Faloutsos. Ratio rules: A new paradigm for fast, quantifiable data mining.
[27]J. Kleinberg, C. Papadimitriou, and P. Raghavan. Segmentation problems. Proceedings of the 30th Annual Symposium on Theory of Computing, ACM. 1998.
[28]J. L. Lin, and M. H. Dunham. Mining association rules: Anti-skew algorithms. Proceedings of the International Conference on Data Engingeering, Orlando, Florida, February 1998.
[29]H. Mannila, H. Toivonen, and A. Verkamo. Efficient algorithm for discovering association rules. AAAI Workshop on Knowledge Discovery in Databases, 1994, pp. 181-192.
[30]田金兰，黄刚，关联规则的发现，计算机世界1999年第20期
[31]Srikant R, Agrawal R. Mining Quantitative Association Rules in Large Relational Tables [J]. Proceedings of ACM SIGMOD International conference on Management of Data, 1996: 1～12.
[32]Brin S, Motwani R, Ullman J et al. Dynamic Itemset Counting and Implication Rules for Market Basket Data [J]. Proceedings of ACM SIGMOD International conference on Management of Data, 1997: 255～264.
[33]Tsur D, Ullman J, Abiteboul S et al. Query Flocks: A Generalization of Association-Rule Mining [J]. Proceedings of ACM SIGMOD International conference on Management of Data, 1998,. 1～12.
[34]Hidber C. Online Association Rule Mining [J]. Proceedings of ACM SIGMOD International conference on Management of Data, 1999:145～156.
[35]Hidber C. Online Association Rule Mining [EB/OL].http://www.icsi.berkeley.edu/～hidber/, 1998.
[36]Hidber C. Carma: update [EB/OL]. http://www.icsi.berkeley.edu/～hidber/, 1999.
[37]Jiawei Han, Micheline Kamber:Data Mining: Concepts and Techniques,2001
[2]李东主编，企业信息化案例，北京大学出版社，P16-34
[3]Charles E．Spurgenon著，张健／袁晓辉／苗沛荣译，以太网技术入门与实现，机械工业出版社P1-2
[4]孙伟纲，企业战略管理与管理驾驶舱，计算机世界，1999-11-15 C19
[5]Hohhof. B. Developing, Information system for Competitiveintelli-gence, Librarytrends. Vol. 43, No. 2, Fall, 1994, P226-238
[6]刘兴，敢问路在何方(对我国企业MIS建设的思考)，IT经理世界，1999年第12期
[7]吕春明，信息化：决策精确之路，中国计算机报，2001-09-06，P67
[8][美]艾森．拉塞尔／保罗．弗里嘉著 张涛／赵陵 译，麦肯锡意识，华夏出版社，P64
[9]Michaele Porter，陈小悦译，竞争战略，华夏出版社
[10]Michaele Porter著，陈小悦译，竞争优势，华夏出版社
[11]尼葛洛庞帝著，胡泳／范海燕译，数字化生存，海南出版社
[12]陈志宏，CRM：顾客关系管理，经济管理，2001年，第9期
[13]丁宁，信息技术革命与企业组织创新，经济管理出版社，2001年
[14]刘首英主编，Dialogue：Listening to the Voice of World's Top Business Leaders，中国发展出版社
[15][美]David Hand等著 张银奎等译，数据挖掘原理，机械工业出版社，2003
[16][美]W．H．Inmon著 王志海 林友芳等译，Building the Data Warehouse，机械工业出版社，2003
[17][美]Alex Berson Ste，构建面向CRM的数据挖掘应用，清华大学出版社，2001
[18]Michael F.S. Chan/Walter W.C. Chung, A Framework to Develop an Enterprise Information Portal for Contract Manufacturing, Int. J. Production Economics 75(2002) 113-126
[19]F. Bezzo/S. Macchietto/C. C. Pantelides, A General Framework for the Integration of Computational Fluid Dynamics and Process Simulation Computers and Chemical Engineerine. 24 (2000)
[20]D.H.R. Price/R. Beach/A.P. Muhlemann/J.A. Sharp/A. Paterson, A System to Support the Enhancement of Strategic Qexibility in Manufacturing Enterprises, European Journal of Operational Research 109(1998) 362-376
[21]Sung HoHa/Sang Chun Park, Application of Data Mining Tools to Hotel Data Mart on the Intranet for Database Marketing, Experts System With Application 15(1998) 1-3
[22]M. Levya/P. Powell/R. Galliers, Assessing Information Systems Strategy Development Frameworks in SMEs, Information & Management 35 (1999) 247-251
[23]R. Agrawal, and J. Shafer. Parallel mining of association rules:Design,Implementaion, and Experience. Technical Repoort FJ10004, IBM
Almaden Research Center, San Jose, CA 95120, Jan. 1996.
[24]S. Brin, R. Motwani, and C. Silverstein. Beyond market baskets:generlizing association rules to correlations. Proceedings of the ACM SIGMOD, 1996. pages 255-276.
[25]S. Brin, R. Motwani, J. D. Ullman, and S. Tsur. Dynamic Itemset counting and implication rules for market basket data. In ACM SIGMOD International Conference On the Management of Data. 1997.
[26]F. Korn, A. Labrinidis, Y. Kotidis, and C. Faloutsos. Ratio rules: A new paradigm for fast, quantifiable data mining.
[27]J. Kleinberg, C. Papadimitriou, and P. Raghavan. Segmentation problems. Proceedings of the 30th Annual Symposium on Theory of Computing, ACM. 1998.
[28]J. L. Lin, and M. H. Dunham. Mining association rules: Anti-skew algorithms. Proceedings of the International Conference on Data Engingeering, Orlando, Florida, February 1998.
[29]H. Mannila, H. Toivonen, and A. Verkamo. Efficient algorithm for discovering association rules. AAAI Workshop on Knowledge Discovery in Databases, 1994, pp. 181-192.
[30]田金兰，黄刚，关联规则的发现，计算机世界1999年第20期
[31]Srikant R, Agrawal R. Mining Quantitative Association Rules in Large Relational Tables [J]. Proceedings of ACM SIGMOD International conference on Management of Data, 1996: 1～12.
[32]Brin S, Motwani R, Ullman J et al. Dynamic Itemset Counting and Implication Rules for Market Basket Data [J]. Proceedings of ACM SIGMOD International conference on Management of Data, 1997: 255～264.
[33]Tsur D, Ullman J, Abiteboul S et al. Query Flocks: A Generalization of Association-Rule Mining [J]. Proceedings of ACM SIGMOD International conference on Management of Data, 1998,. 1～12.
[34]Hidber C. Online Association Rule Mining [J]. Proceedings of ACM SIGMOD International conference on Management of Data, 1999:145～156.
[35]Hidber C. Online Association Rule Mining [EB/OL].http://www.icsi.berkeley.edu/～hidber/, 1998.
[36]Hidber C. Carma: update [EB/OL]. http://www.icsi.berkeley.edu/～hidber/, 1999.
[37]Jiawei Han, Micheline Kamber:Data Mining: Concepts and Techniques,2001
© 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 |