数据挖掘技术在移动通信行业客户关系管理中的应用研究
|
摘要
本论文从数据挖掘技术角度出发，就数据挖掘技术在移动通信客户关系管理(客户关系管理Customer Relation Management，简称CRM)中的应用进行了研究，并在此基础上完成了“客户流失预警”和“客户价值评估”两个具体应用项目的开发。
移动通信市场的竞争实质就是对客户的竞争。这也就提出了CRM实施的必要性和紧迫性，因为CRM的宗旨就是“以客户为中心”。CRM系统要对客户在与企业交互过程中生成的各种数据进行收集、分析，挖掘出隐含在数据中的有用信息，以便给用户提供更有针对性、更优质的服务。这其中必然要注重数据挖掘技术(数据挖掘Data Mining，简称DM)。因此，如何将数据挖掘技术应用于移动通信行业CRM，成为了运营商们最关注的问题之一。
本文首先介绍了DM的算法与技术，特别是本文项目所涉及的决策树算法和Kohonen算法，对分类技术和聚类技术进行了概括；然后介绍了移动通信行业CRM的定义与基本结构，并对与DM技术密切相关的数据仓库技术(数据仓库Data Warehouse，简称DW)作了介绍；接着在DM一般方法论——SEMMA和CRISP-DM的基础上，针对移动通信行业特点，研究了移动通信行业的DM方法论(数据挖掘项目的整体流程)；最后，本文采用了IBM公司的数据挖掘工具(Intelligent Miner，简称IM)和数据仓库工具(DB2)，针对四川移动通信有限责任公司“客户流失预警”和“客户价值评估”两个项目，以项目功能要求为目标，对建立模型的硬、软件环境方案和项目整体流程方面进行设计；在项目的实施过程中，以本文所研究的移动通信行业数据挖掘方法论为基础，对模型的定义、挖掘算法的选择、模型的建立和结果的展现与解释做了详细介绍。
由于目前国内移动通信行业的数据挖掘技术处于起步阶段，许多省市的CRM系统仅仅停留在多维分析、联机分析处理(联机分析处理Online AnalyticalProcessing，简称OLAP)等技术上。本项目的实施推动了数据挖掘技术在移动通信行业CRM的发展和应用，增强了移动运营商的竞争力。
From the point of Data Mining technology, this paper describe the application and research of data mining technology in the industry of mobile telecommunication Customer Relation Management ("CRM" for short). And accomplish two projects: "Customer Churn" and "Customer Value Analyse".
The competition of the telecom market is the competition of the customer. For the kernel ides of CRM method which centers on the customers, We should pay more attention on the necessity and emergency of CRM. CRM system collect, analyse, mining the useful information which is hidden in the data in the process of the activity between the enterprise and the customer. Data Mining ("DM" for short) is one of the important technologies of CRM. So how to use the DM technology in the industry of mobile telecommunication CRM is essential to the operators.
This paper firstly introduce the arithmetic and technology of DM, especially the decision tree and Kohonen, sum up the classification and clustering technology; Secondly, it describe the definition and frame of CRM in the industry of mobile telecommunication and describe Data Warehouse ("DW" for short) as well; Thirdly, based on the SEMMA and CRISP-DM, this paper research the methodology of DM in the industry of mobile telecommunication. At last, aimming at the two Data Mining subjects of Sichuan Mobile Limit Company: "Customer Churn" and "Customer Value Analyse", we give the implementation in detail, including the model definition, the establishment of model and the explain of the result and so
on using the mining tool-Intelligent Miner("IM" for short) and Data Warehouse
tool-DB2.
For the reason of the initial stages of Data Mining of China Telecom industry, This project not only promote the development and applying of Data Mining in mobile communication industry, but also build the competition capability of the telecom operater.
移动通信市场的竞争实质就是对客户的竞争。这也就提出了CRM实施的必要性和紧迫性，因为CRM的宗旨就是“以客户为中心”。CRM系统要对客户在与企业交互过程中生成的各种数据进行收集、分析，挖掘出隐含在数据中的有用信息，以便给用户提供更有针对性、更优质的服务。这其中必然要注重数据挖掘技术(数据挖掘Data Mining，简称DM)。因此，如何将数据挖掘技术应用于移动通信行业CRM，成为了运营商们最关注的问题之一。
本文首先介绍了DM的算法与技术，特别是本文项目所涉及的决策树算法和Kohonen算法，对分类技术和聚类技术进行了概括；然后介绍了移动通信行业CRM的定义与基本结构，并对与DM技术密切相关的数据仓库技术(数据仓库Data Warehouse，简称DW)作了介绍；接着在DM一般方法论——SEMMA和CRISP-DM的基础上，针对移动通信行业特点，研究了移动通信行业的DM方法论(数据挖掘项目的整体流程)；最后，本文采用了IBM公司的数据挖掘工具(Intelligent Miner，简称IM)和数据仓库工具(DB2)，针对四川移动通信有限责任公司“客户流失预警”和“客户价值评估”两个项目，以项目功能要求为目标，对建立模型的硬、软件环境方案和项目整体流程方面进行设计；在项目的实施过程中，以本文所研究的移动通信行业数据挖掘方法论为基础，对模型的定义、挖掘算法的选择、模型的建立和结果的展现与解释做了详细介绍。
由于目前国内移动通信行业的数据挖掘技术处于起步阶段，许多省市的CRM系统仅仅停留在多维分析、联机分析处理(联机分析处理Online AnalyticalProcessing，简称OLAP)等技术上。本项目的实施推动了数据挖掘技术在移动通信行业CRM的发展和应用，增强了移动运营商的竞争力。
From the point of Data Mining technology, this paper describe the application and research of data mining technology in the industry of mobile telecommunication Customer Relation Management ("CRM" for short). And accomplish two projects: "Customer Churn" and "Customer Value Analyse".
The competition of the telecom market is the competition of the customer. For the kernel ides of CRM method which centers on the customers, We should pay more attention on the necessity and emergency of CRM. CRM system collect, analyse, mining the useful information which is hidden in the data in the process of the activity between the enterprise and the customer. Data Mining ("DM" for short) is one of the important technologies of CRM. So how to use the DM technology in the industry of mobile telecommunication CRM is essential to the operators.
This paper firstly introduce the arithmetic and technology of DM, especially the decision tree and Kohonen, sum up the classification and clustering technology; Secondly, it describe the definition and frame of CRM in the industry of mobile telecommunication and describe Data Warehouse ("DW" for short) as well; Thirdly, based on the SEMMA and CRISP-DM, this paper research the methodology of DM in the industry of mobile telecommunication. At last, aimming at the two Data Mining subjects of Sichuan Mobile Limit Company: "Customer Churn" and "Customer Value Analyse", we give the implementation in detail, including the model definition, the establishment of model and the explain of the result and so
on using the mining tool-Intelligent Miner("IM" for short) and Data Warehouse
tool-DB2.
For the reason of the initial stages of Data Mining of China Telecom industry, This project not only promote the development and applying of Data Mining in mobile communication industry, but also build the competition capability of the telecom operater.
引文
[1]陈洪建．客户关系管理(CRM)理论、方法与应用研究．河北工业大学硕士论文．2002，10：11-14
[2]http://www.chinatelecom.com.cn/special/serve/bar4/bar4_4.htm
[3]http://www.crmforum.orglapplyldetail.asp?id=369
[4]Alex Berson，Stephen Smith，Kurt Thearling．[美]构建面向CRM的数据挖掘应用．北京：人民邮电出版社，2001
[5]陈东鹏．数据仓库技术在移动通信领域的应用．电信科学．2001，(5)：34-35
[6]王扶东，朱云龙，薛劲松，李兵．基于数据挖掘的客户关系分析评价系统．东南大学学报(哲学社会科学版)．2002，10(4)：99-102
[7]Quinlan J.R. Decision trees and decision making J IEEE Transaction on Systems, Man, and Cybernetics. 1990, 20(2): 339-346
[8]Quinlan J R. 1986. The effect of noise on concept learning. In: R S Michalske, JG Carbonell, T M Mitchell, eds. Machine Learning: An Artificial Intelligence Approach, Morgan Karufamann
[9]Porter B W, Baress E R, Holte R. 1989. Concept learning and heuristic classification in weak theory domains. Artificial Intelligence, 45(1/2):229-263
[10]Boyen X.Automatic induction of fuzzy decision trees and its application to power system. J Fuzzy Sets and Systems, 1999, 102(1):3-19
[11]R Growth．数据挖掘——构筑企业竞争优势，侯迪等译．西安：西安交通大学出版社，2001：146-186
[12]朱绍文等．决策树采掘技术及发展趋势．计算机工程．2002：28(8)：77-78
[13]J.R. Quinlan. Simplifying decision trees. Interaction, Journal of Man Machine studies. 1987, 27:221-234
[14]L. Brieman, J. Friedman, R.Olshen, and C.Stone. Classification and Regression Trees. Monterey, CA.. Wadswordth International Group, 1984
[15]J.R. Quilan. Induction of decision trees.Machine Learning, 1986, 1:81-106
[16]李广原，陆宇红．基于聚类分析的属性数据挖掘技术．广西师范学院学报(自然科学版)．2002，3(4)：20-23
[17]孙孝萍．基于聚类分析的数据挖掘算法研究．西南石油学院硕士论文．2002，3：26-43
[18]J.MacQueen. Some methods for classification and analysis of multivariate observation. Proc.5th Berkeley Symp. Math Statist, Prob, 1967, 1:281-297
[19]Jianwei Han，Michaline Kamber(加)．数据挖掘概念与技术．范明 孟小峰译．机械工业出版社，2001．8
[20]朱明．数据挖掘．中国科学技术大学出版社，2002．5
[21]Matsuoka K, Ohyz, Kawamoto M. A neural net for blind separation of nonstationary signals[J]. Neural Networks, 1995, 8(3): 411-419
[22]朱明，王俊普．一种聚类学习的新方法．模式识别与人工智能， 2000，9
[23]T.Kohomen.Self-organized formation of topologically correct feature maps.Biological Cybernetics. 1982, 43:59-69
[24]Johan Vesanto, Esa Alhoniemi. Clustering of the Self-Organizing Map[J]. IEEE Transactions on Neural Networks, 2000.11930:586-599
[25]肖怀铁，庄钊文，郭桂荣．基于Kohonen神经网络的飞机目标识别．国防科技大学电子工程学院ATR实验室．现代雷达，1997，19(3)：36-40
[26]T.Kohonen.Self-organized formation of topologically correct feature maps.BiologicalCybernetics, 1982, 43:59-69
[27]R.P.Lippman. An introduction to computing with neural nets IEEE A SSPM agazine, 1987, 4(2): 4-22
[28]Brady J And Davis. Marketing in Transition[J]. The Mckinsey Quarterly, 1993, (2): 17-28
[29]陈旭．CRM综述．计算机应用研究，2001，8：8-10
[30]宋俊德．我国的企业需要CRM．电信技术．2001，5：3-5
[31]Alex Berson, Stephen Smith, Kerr Thearling.Building Data Min2ing Applications for CRM[M]. America:McGraHill Companies, Inc., 2000.
[32]周良，郑洪源，谢强，丁秋林．客户关系管理系统设计研究．计算机应用研究．2002，1：108-109
[33]闫洁．中国电信业的客户关系管理．西南交通大学硕士论文．2003．3
[34]李海明，吴建林，张雷．分析型CRM的软件体系结构．计算机工程与应用． 2002，5：110-112
[35]W．H．Inmon，数据仓库，机械工业出版社，2000
[36]陈燕编著．数据仓库技术及其应用．大连海事大学出版社．2002．3．
[37]廖里，余英泽，吴渔，聂能．数据挖掘和数据仓库及其在电信业中的应用．
重庆邮电学院学报，2002．12(4)：31-34
[38]彭木根编著．数据仓库技术与实现．电子工业出版社，2006：29-32
[39]林宇．数据仓库原理与实践．人民邮电出版社， 2003：231-250
[40]http://dataming.126.com
[41]吕廷杰，尹涛，王琦．客户关系管理与主题分析．人民邮电出版社，2002：108-170
[42]刘莉，徐玉生，马志新．数据挖掘中数据预处理技术综述．甘肃科学学报．2003，15(1)：102-106
[43]刘玉梅．基于数据挖掘技术的电信CRM系统的研究与实现．哈尔滨工程大学硕士论文．2003：44-46
[44]凌立刚．商务智能在电信行业的应用．浙江大学硕士论文，2003：41-42
[45]T. Redman. Data Quality's Management and Technology. Bantam Books, 1992:34-50
[46]R.Wang, V.Storey, and C.Firth. A framework for analysis of data quality research. IEEE Trans. Knowledge and Data Engineering, 1995, 7:623-640
[47]J.H. Friedman. A recursive partitioning decision rule for nonparametric classifiers. IEEETrans. On Comp., 1997, 26:404-408
[48]L.Breiman, J.Friedman, R.Olshen, and C.Stone. Classification and Regression Trees.Monterey, CA: Wadsworth International Group, 1984.
[49]J.R.Quinlan and R.L.Rivest. Inferring decision trees using the minimum description length principle. Information and Computation, 1989:227-248
[50]樊家琨，应用多元分析．河南大学出版社，1993：33-35
[51]www.dwway.com
[2]http://www.chinatelecom.com.cn/special/serve/bar4/bar4_4.htm
[3]http://www.crmforum.orglapplyldetail.asp?id=369
[4]Alex Berson，Stephen Smith，Kurt Thearling．[美]构建面向CRM的数据挖掘应用．北京：人民邮电出版社，2001
[5]陈东鹏．数据仓库技术在移动通信领域的应用．电信科学．2001，(5)：34-35
[6]王扶东，朱云龙，薛劲松，李兵．基于数据挖掘的客户关系分析评价系统．东南大学学报(哲学社会科学版)．2002，10(4)：99-102
[7]Quinlan J.R. Decision trees and decision making J IEEE Transaction on Systems, Man, and Cybernetics. 1990, 20(2): 339-346
[8]Quinlan J R. 1986. The effect of noise on concept learning. In: R S Michalske, JG Carbonell, T M Mitchell, eds. Machine Learning: An Artificial Intelligence Approach, Morgan Karufamann
[9]Porter B W, Baress E R, Holte R. 1989. Concept learning and heuristic classification in weak theory domains. Artificial Intelligence, 45(1/2):229-263
[10]Boyen X.Automatic induction of fuzzy decision trees and its application to power system. J Fuzzy Sets and Systems, 1999, 102(1):3-19
[11]R Growth．数据挖掘——构筑企业竞争优势，侯迪等译．西安：西安交通大学出版社，2001：146-186
[12]朱绍文等．决策树采掘技术及发展趋势．计算机工程．2002：28(8)：77-78
[13]J.R. Quinlan. Simplifying decision trees. Interaction, Journal of Man Machine studies. 1987, 27:221-234
[14]L. Brieman, J. Friedman, R.Olshen, and C.Stone. Classification and Regression Trees. Monterey, CA.. Wadswordth International Group, 1984
[15]J.R. Quilan. Induction of decision trees.Machine Learning, 1986, 1:81-106
[16]李广原，陆宇红．基于聚类分析的属性数据挖掘技术．广西师范学院学报(自然科学版)．2002，3(4)：20-23
[17]孙孝萍．基于聚类分析的数据挖掘算法研究．西南石油学院硕士论文．2002，3：26-43
[18]J.MacQueen. Some methods for classification and analysis of multivariate observation. Proc.5th Berkeley Symp. Math Statist, Prob, 1967, 1:281-297
[19]Jianwei Han，Michaline Kamber(加)．数据挖掘概念与技术．范明 孟小峰译．机械工业出版社，2001．8
[20]朱明．数据挖掘．中国科学技术大学出版社，2002．5
[21]Matsuoka K, Ohyz, Kawamoto M. A neural net for blind separation of nonstationary signals[J]. Neural Networks, 1995, 8(3): 411-419
[22]朱明，王俊普．一种聚类学习的新方法．模式识别与人工智能， 2000，9
[23]T.Kohomen.Self-organized formation of topologically correct feature maps.Biological Cybernetics. 1982, 43:59-69
[24]Johan Vesanto, Esa Alhoniemi. Clustering of the Self-Organizing Map[J]. IEEE Transactions on Neural Networks, 2000.11930:586-599
[25]肖怀铁，庄钊文，郭桂荣．基于Kohonen神经网络的飞机目标识别．国防科技大学电子工程学院ATR实验室．现代雷达，1997，19(3)：36-40
[26]T.Kohonen.Self-organized formation of topologically correct feature maps.BiologicalCybernetics, 1982, 43:59-69
[27]R.P.Lippman. An introduction to computing with neural nets IEEE A SSPM agazine, 1987, 4(2): 4-22
[28]Brady J And Davis. Marketing in Transition[J]. The Mckinsey Quarterly, 1993, (2): 17-28
[29]陈旭．CRM综述．计算机应用研究，2001，8：8-10
[30]宋俊德．我国的企业需要CRM．电信技术．2001，5：3-5
[31]Alex Berson, Stephen Smith, Kerr Thearling.Building Data Min2ing Applications for CRM[M]. America:McGraHill Companies, Inc., 2000.
[32]周良，郑洪源，谢强，丁秋林．客户关系管理系统设计研究．计算机应用研究．2002，1：108-109
[33]闫洁．中国电信业的客户关系管理．西南交通大学硕士论文．2003．3
[34]李海明，吴建林，张雷．分析型CRM的软件体系结构．计算机工程与应用． 2002，5：110-112
[35]W．H．Inmon，数据仓库，机械工业出版社，2000
[36]陈燕编著．数据仓库技术及其应用．大连海事大学出版社．2002．3．
[37]廖里，余英泽，吴渔，聂能．数据挖掘和数据仓库及其在电信业中的应用．
重庆邮电学院学报，2002．12(4)：31-34
[38]彭木根编著．数据仓库技术与实现．电子工业出版社，2006：29-32
[39]林宇．数据仓库原理与实践．人民邮电出版社， 2003：231-250
[40]http://dataming.126.com
[41]吕廷杰，尹涛，王琦．客户关系管理与主题分析．人民邮电出版社，2002：108-170
[42]刘莉，徐玉生，马志新．数据挖掘中数据预处理技术综述．甘肃科学学报．2003，15(1)：102-106
[43]刘玉梅．基于数据挖掘技术的电信CRM系统的研究与实现．哈尔滨工程大学硕士论文．2003：44-46
[44]凌立刚．商务智能在电信行业的应用．浙江大学硕士论文，2003：41-42
[45]T. Redman. Data Quality's Management and Technology. Bantam Books, 1992:34-50
[46]R.Wang, V.Storey, and C.Firth. A framework for analysis of data quality research. IEEE Trans. Knowledge and Data Engineering, 1995, 7:623-640
[47]J.H. Friedman. A recursive partitioning decision rule for nonparametric classifiers. IEEETrans. On Comp., 1997, 26:404-408
[48]L.Breiman, J.Friedman, R.Olshen, and C.Stone. Classification and Regression Trees.Monterey, CA: Wadsworth International Group, 1984.
[49]J.R.Quinlan and R.L.Rivest. Inferring decision trees using the minimum description length principle. Information and Computation, 1989:227-248
[50]樊家琨，应用多元分析．河南大学出版社，1993：33-35
[51]www.dwway.com