数据挖掘技术在企业CRM系统中的研究与应用

发布时间：2018-06-22 12:48

本文选题：客户关系管理 + 数据挖掘　；参考：《浙江理工大学》2015年硕士论文

【摘要】：伴随着互联网技术的飞速发展，企业之间的竞争日益激烈，传统的营销模式受到严重的挑战，企业需要并得以借助先进的管理思想以及先进的技术手段，去充分了解和掌握客户信息，，发现与挖掘潜在市场机会，规避风险，提高客户忠诚度与满意度。企业对于客户信息的获取以及如何使这些信息发挥最大的效用，是企业在自身发展中的一个重要工作，这就必须有一个专门的系统来收集客户信息，并加以分析和利用。客户关系管理(CustomerRelationship Management,CRM)系统正是在此环境下应运而生，它将利用企业各种资源，通过客户的管理过程，分析出企业优劣，提高企业的竞争力，使企业在现如今激烈的竞争环境中处于不败之地。同时，随着企业业务的不断扩大，企业里的数据信息正以指数级的速度在迅速增长，特别是客户信息中不仅包含着有价值的部分，同时也包含很多冗余的信息，要想在这海量的信息中准确的找到企业需要的资源，从中提取出客户关系图，也是一个难点。因此，需要在构建CRM系统的同时应用现如今正飞速发展的数据挖掘技术，将两者结合起来，才能搭建出一个完整的CRM系统。本文结合浙江中烟工业公司（以下简称浙江中烟）CRM系统，分析了企业中海量客户数据处理的方法，并对客户进行划分，提取客户关系图以及检测客户忠诚度等。浙江中烟CRM系统对企业的发展起到了至关重要的作用。主要研究有以下几个方面： 1)以浙江中烟为例，结合大型集团企业CRM系统的需求分析，了解CRM系统的特性、架构和运行机制，分析企业CRM系统的可行性以及数据挖掘技术如何的应用于系统当中。 2)构建客户分类决策树，提取客户的行为记录，如反馈、来访、活动交流等，分析客户的忠诚度状态；建立客户忠诚度预警模型，根据CLV曲线分析，得出客户在某个时间点或时间范围内的忠诚度，并据此提出提升客户忠诚度的方案。 3)提出了一种基于改进的FP-Growth算法的客户关系图提取方法，使得企业能够很清晰的看出与客户之间的关系，找出存在的不足加以改进，使企业决策者和服务人员做出相应的决策行为，改善客户关系，提高客户的忠诚度。 4)提出了一种基于TFIDF算法同义替换和相邻合并的文本挖掘技术，可以降低服务器压力，使服务人员可以更快更准确的从知识库中寻找出相关信息，满足客户的需求。
[Abstract]:With the rapid development of Internet technology, the competition between enterprises is increasingly fierce, the traditional marketing model is seriously challenged, enterprises need and can use advanced management ideas and advanced technical means. To fully understand and master customer information, identify and tap potential market opportunities, avoid risks, and improve customer loyalty and satisfaction. It is an important work for enterprises to obtain customer information and how to make it play a most important role in their own development. So it is necessary to have a special system to collect customer information and analyze and utilize it. The customer relationship Management (CRM) system emerges as the times require in this environment, it will use all kinds of resources of the enterprise, through the customer management process, analyze the advantages and disadvantages of the enterprise, improve the competitiveness of the enterprise. So that enterprises in today's fierce competition environment in an invincible position. At the same time, with the continuous expansion of enterprise business, the data information in the enterprise is growing exponentially, especially the customer information not only contains valuable parts, but also contains a lot of redundant information. It is also a difficult point to find the resources needed by the enterprise and extract the customer relationship diagram from the massive information. Therefore, it is necessary to build a complete CRM system by applying the data mining technology which is developing at full speed in order to build a complete CRM system. Based on the CRM system of Zhejiang Zhongyan Industrial Corporation (hereinafter referred to as "Zhejiang Zhongyan"), this paper analyzes the methods of processing massive customer data in enterprises, and divides customers, extracts customer relationship diagrams and detects customer loyalty. Zhejiang Zhongyan CRM system plays a vital role in the development of enterprises. The main research has the following several aspects: 1) take Zhejiang Zhongyan as an example, combined with the large-scale group enterprise CRM system demand analysis, understands the CRM system characteristic, the structure and the movement mechanism, Analyze the feasibility of CRM system and how to apply data mining technology in the system. 2) construct customer classification decision tree, extract customer behavior record, such as feedback, visit, activity exchange, etc. Based on the analysis of CLV curve, the customer loyalty in a certain time or time range can be obtained. Based on the improved FP-Growth algorithm, a customer relationship graph extraction method is proposed, which enables enterprises to clearly see the relationship with customers. To find out the shortcomings to improve, so that the enterprise decision makers and service personnel to make the corresponding decision-making behavior, improve customer relations, (4) A text mining technology based on synonymous substitution and adjacent merging of TFIDF algorithm is proposed, which can reduce the server pressure and make the service staff find the relevant information from the knowledge base more quickly and accurately. Meet the needs of customers.
【学位授予单位】：浙江理工大学
【学位级别】：硕士
【学位授予年份】：2015
【分类号】：TP311.13

【参考文献】