数据挖掘中的分类技术在我国保险业CRM的研究应用
|
摘要
随着计算机技术和经济的飞速发展,全球各行服务业的数据库不仅在数量上及规模上也发生了翻天覆地的变化,数据挖掘正是在这样的应用需求背景下产生并迅速发展起来的一个重要的研究领域,数据挖掘技术作为多学科的集成,已成为最重要的信息处理技术和方法。
客户关系管理作为一种以客户为中心的经营策略,可以很好完成客户的获取与保持,为经营者提供决策支持,而数据挖掘作为一种分析工具,可以应用在客户关系管理中的大量数据分析和客户价值挖掘中。因此,在客户关系管理中构建有效的数据挖掘应用,研究有助于提高客户关系管理中的决策支持功能的数据挖掘技术和理论,是非常有意义的一个课题。
本文介绍了数据挖掘技术中的分类技术在我国保险业客户关系管理中的应用,主要采用了决策树和神经网络算法对保险业客户关系管理中的客户细分、客户流失等方面进行分析,并在研究了相关客户关系管理体系结构的基础上,通过实验来构建一种基于weka平台的保险业客户关系管理系统,并给出部分挖掘结果。
在客户群体细分中,提出了一种基于ID3算法的改进算法,一般来说,准确迅速的选择属性,使计算量减少,就可加快算法生成树的速度。选择年龄、驾龄、平均赔付率等5个属性作为条件属性,分别运用ID3算法及其改进算法选出决策属性,得出决策树,以细分具有哪些特征客户具有高的风险。
客户流失分析采用了分类算法中的神经网络算法,选择年龄、文化程度、年收入、工作地区、职业、险种等属性作为相关客户数据指标,将这些指标输入遗传神经网络,经过训练得出客户流失预测模型,通过预测模型来指导决策者是否对特定客户群采取必要措施来降低流失率以及对哪些客户采取挽留措施。
With the computer technology and the rapid economic development, the global database of all line services have changed tremendously in number and scale .Data Mining is in the context of this application needs to generate and rapidly becomes an important area of research. Data Mining technology as a multi-disciplinary integration, has become the most important information-processing techniques and methods.
As a kind of management strategy of customer centered, Customer Relationship Man- agement (CRM) can retain and obtain customers well. Data Mining, as a tool of analyses, can be used to analyze mass data and mine customer values in CRM.Therefore,building effective DM applications in CRM and researching which can improve decision-making support functions of data mining techniques and theory is a topic of great significance.
In this thesis, I research the application of DM’s Classification technology in the insurance CRM of China, mainly using decision-tree and neural network algorithms to analyze the particularly customers distinguish, and so on. In the end, After studying the relevant CRM system, I build a new CRM system through the experiments based on weka platform and give part of the mining results.
In customers distinguish, I put forward an improved algorithm based on ID3 algorithm. In general, in order to accelerate the speed of spanning tree algorithm, we must choose properties accuratelly and quickly. Finally, I choose age, driving experience, the average payment rate of 5 attributes as condition attributes and select decision-making attributes in order to obtain decision tree, which features a detailed breakdown with a client with a high risk.
In Customers loss analysis, I use neural network algorithm and choose the age, education level, income, work area, occupation, insurance and other properties as the relevant customer data indicators. These indicators will enter the genetic neural network and be trained to identify our customer loss prediction model which can guide decision-makers whether a specific customer base to take the necessary measures to reduce loss and retention measures to be taken about which kinds of customer.
客户关系管理作为一种以客户为中心的经营策略,可以很好完成客户的获取与保持,为经营者提供决策支持,而数据挖掘作为一种分析工具,可以应用在客户关系管理中的大量数据分析和客户价值挖掘中。因此,在客户关系管理中构建有效的数据挖掘应用,研究有助于提高客户关系管理中的决策支持功能的数据挖掘技术和理论,是非常有意义的一个课题。
本文介绍了数据挖掘技术中的分类技术在我国保险业客户关系管理中的应用,主要采用了决策树和神经网络算法对保险业客户关系管理中的客户细分、客户流失等方面进行分析,并在研究了相关客户关系管理体系结构的基础上,通过实验来构建一种基于weka平台的保险业客户关系管理系统,并给出部分挖掘结果。
在客户群体细分中,提出了一种基于ID3算法的改进算法,一般来说,准确迅速的选择属性,使计算量减少,就可加快算法生成树的速度。选择年龄、驾龄、平均赔付率等5个属性作为条件属性,分别运用ID3算法及其改进算法选出决策属性,得出决策树,以细分具有哪些特征客户具有高的风险。
客户流失分析采用了分类算法中的神经网络算法,选择年龄、文化程度、年收入、工作地区、职业、险种等属性作为相关客户数据指标,将这些指标输入遗传神经网络,经过训练得出客户流失预测模型,通过预测模型来指导决策者是否对特定客户群采取必要措施来降低流失率以及对哪些客户采取挽留措施。
With the computer technology and the rapid economic development, the global database of all line services have changed tremendously in number and scale .Data Mining is in the context of this application needs to generate and rapidly becomes an important area of research. Data Mining technology as a multi-disciplinary integration, has become the most important information-processing techniques and methods.
As a kind of management strategy of customer centered, Customer Relationship Man- agement (CRM) can retain and obtain customers well. Data Mining, as a tool of analyses, can be used to analyze mass data and mine customer values in CRM.Therefore,building effective DM applications in CRM and researching which can improve decision-making support functions of data mining techniques and theory is a topic of great significance.
In this thesis, I research the application of DM’s Classification technology in the insurance CRM of China, mainly using decision-tree and neural network algorithms to analyze the particularly customers distinguish, and so on. In the end, After studying the relevant CRM system, I build a new CRM system through the experiments based on weka platform and give part of the mining results.
In customers distinguish, I put forward an improved algorithm based on ID3 algorithm. In general, in order to accelerate the speed of spanning tree algorithm, we must choose properties accuratelly and quickly. Finally, I choose age, driving experience, the average payment rate of 5 attributes as condition attributes and select decision-making attributes in order to obtain decision tree, which features a detailed breakdown with a client with a high risk.
In Customers loss analysis, I use neural network algorithm and choose the age, education level, income, work area, occupation, insurance and other properties as the relevant customer data indicators. These indicators will enter the genetic neural network and be trained to identify our customer loss prediction model which can guide decision-makers whether a specific customer base to take the necessary measures to reduce loss and retention measures to be taken about which kinds of customer.
引文
[1]何荣勤. CRM原理.设计.实践(第2版).北京:电子工业出版社,2006.
[2]叶开.中国CRM最佳实务.北京:电子工业出版社2005.4?,p1-10.
[3]管政,魏冠明.中国企业CRM实施.北京:人民邮电出版社2003.3, p3-10.
[4] Ian H.Witten, Eibe Frank著,董琳,邱泉等译.数据挖掘实用机器学习技术.北京,机械工业出版社2006.2
[5]邱晓蕾,陆黎明.数据挖掘技术在保险业中的应用[J].福建电脑,2005(9),134-135.
[6]唐洪浪,桂现才.数据挖掘技术在客户关系管理中的应用[J].湛江师范学院学报,2004(12),124-130.
[7]俞明春,黄江燕.构建基于数据挖掘技术的客户关系管理系统[J].科技广场,2005(12)
[8]梁晶.CRM在中资保险业中的应用及其风险分析. [硕士学位论文].南京师范大学,2005年4月,3-15.
[11]张震宇.数据挖掘技术在保险业CRM中的应用研究. [硕士学位论文].重庆大学.2004, 30-45.
[12]王辉.我国保险业客户关系管理研究. [硕士学位论文].天津财经大学.2006年6月, 25-45.
[13]余林慧.保险公司客户关系管理研究. [硕士学位论文].武汉大学.2005年5月, 33-45.
[14]张清.平安保险公司客户关系管理(CRM)研究.[硕士学位论文].西南交通大学.2002年5月, 12-41.
[15]姜金贵.数据挖掘分类算法在CRM中的研究.[硕士学位论文].哈尔滨工程大学.2005年2月, 22-31.
[16]叶开.中国CRM最佳实务.北京:电子工业出版社,2005年4月, 12-16.
[17]朱明.数据挖掘CRM中的应用.安徽:中国科学与技术大学出版社,2002, 4-10.
[18]周洁如,庄晖.现代客户关系管理.上海:上海交通大学出版社.2008,1-20.
[19]林宇等.数据仓库原理与实践,北京:人民邮电出版社,2004
[20] James,D.Better together marketing research and CRM Marketing News 2002, 36(10):15-16.
[21] Wolfgang Ulaga. Customer Value in Business Markets: An Agenda for Inquiry. Industrial marketing Management,No.:30,2001,P315-319.
[22]王洪艳.基于聚类的数据挖掘技术在CRM中的研究与应用. [硕士学位论文].武汉:武汉大学.2005年5月.
[23] Jiawei Han、Micheline Kamber.数据挖掘概念与技术[M].机械工业出版社, 2007.7.,69-269
[24] GTS Ho. An intelligent information infrastructure to support the streamlining of integrated logistics workflow[J]. Expert Systems, 2004, 21(3):123-131.
[25]毛国君,段立娟,王实等.数据挖掘原理与算法[M].北京:清华大学出版社, 2005.,1-33.
[26]柳林,涂光平,杨峰.基于决策树的数据挖掘方法在CRM中的应用研究[J].计算技术与自动化.2006,25(1),67-69.
[27]王荣.分类技术及其在客户关系管理中的应用. [硕士学位论文].杭州:浙江大学.2006年3月,45-48.
[28]罗丹.基于CRM与数据挖掘的客户分类研究.[硕士学位论文].阜新:辽宁工程技术大学,2004年12月,36-43.
[29] R .Agrawal,Imielinski.T,Swami.A.,Mining Association rules between Sets of Items in large Databases[C]. In Proc.1993 ACM-SIGMOD Int.Conf.Management of Data(SIGMOD’93) 207-216, Washington D.C,July 1993.
[30] R.Agrawal, R.Srikant.Fast Algorithms for Mining Association Rules[C]. Proceedings of the 20th International Conference on Very Large Databases (VLDB'94), Santiago, Chile, 1994:487-499.
[31] James Cheng, Yiping Ke, Wilfred Ng. Maintaining Frequent Itemsets over High-Speed Data Streams[C]. In PAKDD 2006, pp.462-467, 2006.
[32] H..D.K Moonesinghe, et al. Frequent Closed Itemset Mining Using Prefix Graphs with An Efficient Flow-Based Pruning Strategy[C]. In ICDM 2006, pp.426-435, 2006.
[33] Gaurav Pandey, Michael Steinbach, Rohit GuPta, Tushar Garg, and VIPin Kumar. Association Analysis-based Transformations for Protein Interaction Networks: A Function Prediction Case Study[C]. In SIGKDD’07, San Jose, California, 2007.
[34] Chun-Kit Chui, Ben Kao, Edward Hung. Mining frequent itemsets from uncertain data[C]. PAKDD 2007, LNAI 4426, pp.47-58, 2007.
[35] Yuefeng Li, Wangzhong Yang, et al. Multi-Tier granule mining for representations of multidimention association rules[C]. In Porc. Of the 6th IEEE Interactional Conference on Data Mining(ICDM’06), pp. 953-958, 2006.
[36] Laure Berti-Equille. Quality-aware association rule mining[C]. In PAKDD 2006, pp.440-449, 2006.
[37] Yun Sing Koh, Nathan Rountree, Richard O’Keefe. Mining interesting imperfectly sporadic rules. InProc[C]. Of 10th Pacific-Asia Conference on Advances in Knowledge Discovery and Data Mining(PAKDD’06), LNCS 3918, pp.473-482, 2006.
[38] Shin-Yuan Hung,David C Yen. Hsiu-Yu Wang. Appyling data mining telecom chum Management[J].Expert Systems with Applications,2005.9:l-10
[39] Shi Zhong,Taghi Khoshgoftaar,Naeem Seliya.Evaluating Clustering Techniques for Network Intrusi on Detection[M].Florida Atlantic University, 2002:78-93.
[40] S.Zanero S.M.Savaresi. Unsupervised Learning Techniques for an Intrusion Detection System [A], Proceedings of the 2004 ACM Symposium on Applied Computing, Nicosia, Cyprus, 2004.
[41] K.Wang, S.J.Stolfo. Anomalous Payload-based Network Intrusion Detection[C], In Proceedings of the Seventh International Symposium on Recent Advances in Intrusion Detection (RAID),2004.
[42] An Introduction to Intrusion Detection and Assessment. http://www.icsa.com.
[43] Yao X. Evolving artificial neural networks[A]. Proceedings of the IEEE. 1997.89(9): 1423-1447.
[44]刘鹏,姚正,尹俊杰.一种有效的C4.5改进模型[J].清华大学学报(自然科学版),2006(S1):99621001,997-1001.
[45] P.Datta,B.Masand.D.R.Mani,B.Li. Automated cellular modeling and Prediction on a larger scale[J]. Artificial Intelligence Review 2000,485-502.
[2]叶开.中国CRM最佳实务.北京:电子工业出版社2005.4?,p1-10.
[3]管政,魏冠明.中国企业CRM实施.北京:人民邮电出版社2003.3, p3-10.
[4] Ian H.Witten, Eibe Frank著,董琳,邱泉等译.数据挖掘实用机器学习技术.北京,机械工业出版社2006.2
[5]邱晓蕾,陆黎明.数据挖掘技术在保险业中的应用[J].福建电脑,2005(9),134-135.
[6]唐洪浪,桂现才.数据挖掘技术在客户关系管理中的应用[J].湛江师范学院学报,2004(12),124-130.
[7]俞明春,黄江燕.构建基于数据挖掘技术的客户关系管理系统[J].科技广场,2005(12)
[8]梁晶.CRM在中资保险业中的应用及其风险分析. [硕士学位论文].南京师范大学,2005年4月,3-15.
[11]张震宇.数据挖掘技术在保险业CRM中的应用研究. [硕士学位论文].重庆大学.2004, 30-45.
[12]王辉.我国保险业客户关系管理研究. [硕士学位论文].天津财经大学.2006年6月, 25-45.
[13]余林慧.保险公司客户关系管理研究. [硕士学位论文].武汉大学.2005年5月, 33-45.
[14]张清.平安保险公司客户关系管理(CRM)研究.[硕士学位论文].西南交通大学.2002年5月, 12-41.
[15]姜金贵.数据挖掘分类算法在CRM中的研究.[硕士学位论文].哈尔滨工程大学.2005年2月, 22-31.
[16]叶开.中国CRM最佳实务.北京:电子工业出版社,2005年4月, 12-16.
[17]朱明.数据挖掘CRM中的应用.安徽:中国科学与技术大学出版社,2002, 4-10.
[18]周洁如,庄晖.现代客户关系管理.上海:上海交通大学出版社.2008,1-20.
[19]林宇等.数据仓库原理与实践,北京:人民邮电出版社,2004
[20] James,D.Better together marketing research and CRM Marketing News 2002, 36(10):15-16.
[21] Wolfgang Ulaga. Customer Value in Business Markets: An Agenda for Inquiry. Industrial marketing Management,No.:30,2001,P315-319.
[22]王洪艳.基于聚类的数据挖掘技术在CRM中的研究与应用. [硕士学位论文].武汉:武汉大学.2005年5月.
[23] Jiawei Han、Micheline Kamber.数据挖掘概念与技术[M].机械工业出版社, 2007.7.,69-269
[24] GTS Ho. An intelligent information infrastructure to support the streamlining of integrated logistics workflow[J]. Expert Systems, 2004, 21(3):123-131.
[25]毛国君,段立娟,王实等.数据挖掘原理与算法[M].北京:清华大学出版社, 2005.,1-33.
[26]柳林,涂光平,杨峰.基于决策树的数据挖掘方法在CRM中的应用研究[J].计算技术与自动化.2006,25(1),67-69.
[27]王荣.分类技术及其在客户关系管理中的应用. [硕士学位论文].杭州:浙江大学.2006年3月,45-48.
[28]罗丹.基于CRM与数据挖掘的客户分类研究.[硕士学位论文].阜新:辽宁工程技术大学,2004年12月,36-43.
[29] R .Agrawal,Imielinski.T,Swami.A.,Mining Association rules between Sets of Items in large Databases[C]. In Proc.1993 ACM-SIGMOD Int.Conf.Management of Data(SIGMOD’93) 207-216, Washington D.C,July 1993.
[30] R.Agrawal, R.Srikant.Fast Algorithms for Mining Association Rules[C]. Proceedings of the 20th International Conference on Very Large Databases (VLDB'94), Santiago, Chile, 1994:487-499.
[31] James Cheng, Yiping Ke, Wilfred Ng. Maintaining Frequent Itemsets over High-Speed Data Streams[C]. In PAKDD 2006, pp.462-467, 2006.
[32] H..D.K Moonesinghe, et al. Frequent Closed Itemset Mining Using Prefix Graphs with An Efficient Flow-Based Pruning Strategy[C]. In ICDM 2006, pp.426-435, 2006.
[33] Gaurav Pandey, Michael Steinbach, Rohit GuPta, Tushar Garg, and VIPin Kumar. Association Analysis-based Transformations for Protein Interaction Networks: A Function Prediction Case Study[C]. In SIGKDD’07, San Jose, California, 2007.
[34] Chun-Kit Chui, Ben Kao, Edward Hung. Mining frequent itemsets from uncertain data[C]. PAKDD 2007, LNAI 4426, pp.47-58, 2007.
[35] Yuefeng Li, Wangzhong Yang, et al. Multi-Tier granule mining for representations of multidimention association rules[C]. In Porc. Of the 6th IEEE Interactional Conference on Data Mining(ICDM’06), pp. 953-958, 2006.
[36] Laure Berti-Equille. Quality-aware association rule mining[C]. In PAKDD 2006, pp.440-449, 2006.
[37] Yun Sing Koh, Nathan Rountree, Richard O’Keefe. Mining interesting imperfectly sporadic rules. InProc[C]. Of 10th Pacific-Asia Conference on Advances in Knowledge Discovery and Data Mining(PAKDD’06), LNCS 3918, pp.473-482, 2006.
[38] Shin-Yuan Hung,David C Yen. Hsiu-Yu Wang. Appyling data mining telecom chum Management[J].Expert Systems with Applications,2005.9:l-10
[39] Shi Zhong,Taghi Khoshgoftaar,Naeem Seliya.Evaluating Clustering Techniques for Network Intrusi on Detection[M].Florida Atlantic University, 2002:78-93.
[40] S.Zanero S.M.Savaresi. Unsupervised Learning Techniques for an Intrusion Detection System [A], Proceedings of the 2004 ACM Symposium on Applied Computing, Nicosia, Cyprus, 2004.
[41] K.Wang, S.J.Stolfo. Anomalous Payload-based Network Intrusion Detection[C], In Proceedings of the Seventh International Symposium on Recent Advances in Intrusion Detection (RAID),2004.
[42] An Introduction to Intrusion Detection and Assessment. http://www.icsa.com.
[43] Yao X. Evolving artificial neural networks[A]. Proceedings of the IEEE. 1997.89(9): 1423-1447.
[44]刘鹏,姚正,尹俊杰.一种有效的C4.5改进模型[J].清华大学学报(自然科学版),2006(S1):99621001,997-1001.
[45] P.Datta,B.Masand.D.R.Mani,B.Li. Automated cellular modeling and Prediction on a larger scale[J]. Artificial Intelligence Review 2000,485-502.