【作者】 周露；

【导师】 徐绪松；

【作者基本信息】 武汉大学 ， 技术经济及管理， 2004， 硕士

【摘要】 随着知识经济时代的到来和经济全球化的趋势，企业间的竞争日益激烈，客户成为企业至关重要的成功因素和利润来源。企业的资本、技术、商业模式越来越容易模仿，而与客户的关系是建立在长期积累客户交易资料、深入熟悉客户需求的基础上的，无法复制，因此，把握客户是电子商务时代构建企业核心竞争力的关键。Internet的普及，大大加强了客户的知识力量，也有人说经济全球化导致的激烈市场竞争使这些客户变得越来越难以驾驭，不论何种原因，有一点是肯定的，即信息和技术资源在全球范围内的大幅度重新分配使得客户在与企业的权力争夺战中，逐渐掌握了越来越大的主动权和选择权。面对“电子化客户”的挑战，许多企业搜集和存储了关于客户、供应商和商业伙伴的宝贵数据，由于缺乏发现隐含在数据中的有用的信息的能力，所以这些企业无法将数据转化为知识。因此企业应该利用先进的数据挖掘技术分析客户的各种背景数据和过去的交易行为数据，从中获取知识，牢牢锁定客户。 本文研究了电子化客户关系管理(eCRM)相关理论，介绍了数据仓库基本理论和数据挖掘技术的特点、内涵、工作流程、建模、分析技术和挖掘工具等。并讨论了在电子化客户关系管理系统的实现中如何融入数据仓库机制和数据挖掘技术的几个关键问题。在研究了电子化客户关系管理系统的功能模块、数据挖掘步骤及如何获取客户数据从而构建数据仓库的基础上，重点解决了客户关系管理系统中的数据分析方法，利用数据挖掘技术中BP神经网络算法、决策树的贪心算法和聚类分析中的K-平均算法构建合适的预测分类模型，解决了获取客户、客户保持和个性化客户服务的算法实现问题，从而实现数据挖掘技术在电子化客户关系管理中的应用。 本文具体内容如下： 第一章——电子化客户关系管理(eCRM)。研究了eCRM的起源、内涵和体系结构等理论基础，并提出了客户满意、客户生命周期、客户收益递增、客户锁定、客户保持等eCRM基本原则。 第二章——数据挖掘技术。介绍了数据仓库技术以及数据挖掘的特点、内涵、工作流程、建模、分析技术和数据挖掘工具等。 第三章——数据挖掘技术在eCRM应用中的关键问题。介绍了eCRM的系统功能模块和基于数据挖掘技术的eCRM系统应用模型，提出了在eCRM系统进行数据挖掘的步骤，并且论述了从网上获取客户数据的方法，构建了网上书店客户数据库表结构。 第四章—案例:网上书店电子化客户关系管理系统中的数据挖掘算法设计。提出了获取新客户的神经网络算法、保持老客户的决策树算法、提高客户满意度的聚类算法。 论文具有以下特点及创新: (l)在导师徐绪松教授的《复杂科学·资本市场·项目评价》一书中提出的定性定量结合的理论框架基础上，分别运用了人工神经网络、决策树技术、聚类技术等有效的解决了电子化客户关系管理中的客户获取、客户保持和个性化客户服务的问题。 (2)在研究了营销理论、客户关系、企业经营活动和客户价值的基础上，归纳出了电子化客户关系管理中的客户满意、客户保持、客户收益递增、客户锁定等原则，归纳出客户生命周期模型。更多还原

【Abstract】 With the advent of knowledge economic epic and the trend of global economy, the more and more intense competition among enterprises, winning customers become the important key to success and profit source. Because the enterprise capital, technology and business model are easier and easier to simulate, while the relationship with customers is built on accumulating customer transaction information and deep familiarization with customer requirement, which cannot be replicated, winning customer is the key to build enterprise kernel competition in electronic business age. The wide application of Internet greatly improves customer’s knowledge power. Also some people believe that the intense marketing competition caused by economic globalization makes the customers more and more difficult to tract. Whatever reasons, we are sure that the global redistribution of information and technology resource makes the customers have more and more choices caused by the competition among enterprises. Facing the challenge of "electronic customer", many enterprises collect and store valuable data of customers, vendors and business partners. Because of lacking the ability to find out the implicit information in the valuable data, these enterprises cannot change the data into knowledge. Thus enterprises should exploit advanced data mining technology to analyze the background data and past transaction data, get knowledge and win the customer.This paper studies related theories of electronic customer relationship management (eCRM), combines the data mining mechanism and technology into the implementation of eCRM. Based on the study of function models of eCRM system, this paper solves the data analyze methods, exploits data mining technology algorithms to construct proper customer life cycle model, predicts future trend, and implements the application of data mining technology in eCRM.The paper is organized as follows:Chapter 1, eCRM, studies the origin, content and architecture of eCRM, proposes the basic principles of customer satisfaction, customer life cycle, increasing customer revenue, customer targeted, customer keep, and so on.Chapter 2, data mining technology, presents data ware technology and its characteristic, mechanism, process, model, analyze technology and mining tools, and so on.Chapter 3, the key questions of data mining technology in the application of eCRM, introduces the system function models of eCRM, and the system application model based on data mining technology, proposes the data mining procedure of eCRM system., and presents the method to get customer data from Internet, and builds the database table structure of Internet bookstore.Chapter 4, example, the data mining algorithm design of a Internet bookstore eCRM, presents the neural network algorithm of getting new customers, decision tree algorithm to keep old customers, the clustering algorithm to improve customer satisfaction.This paper has the following innovations.(1) Combine both quality and quantity methods, exploit "the latest technology in both the social science and natural science, propose the data mining technology application of three algorithms to eCRM of Internet bookstore: neural network algorithm of wining new customers, decision algorithm to keep old customers, and clustering algorithm to improve customer satisfaction.(2) summarize the principles of customer satisfaction, increasing customer revenue, customer targeted, customer keep, construct customer life cycle model.更多还原