网购客户流失的实证分析

发布时间：2018-08-28 09:41

【摘要】：近年来,随着生产力的不断提高,信息技术的大力发展,互联网成为了当今社会的重要战略资源。伴随着互联网时代的到来,企业的商务环境也发生了翻天覆地的变化,电子商务平台得以构建。电子商务模式的简单、快捷和方便性吸引了大量客户的目光。许多客户纷纷转向了这个新兴的市场,加入了网购的大军。在电子商务平台上,企业仅仅通过吸引新客户,提高市场份额来赢得这场商战的胜利是远远不够的。电商企业还必须做好客户流失的防御工作,解决好客户“进”和“出”的问题,才能达到电商企业对客户的有效管理的目的。目前关于客户流失的研究更多的还是关注传统企业,而较少去涉及B2C平台或C2C平台这种新型商务环境中的企业。在这个网购已经成为一种生活方式的时代,传统研究还很难在电子商务领域得到运用。因此,本文将传统的客户流失预测模型与电子商务模式相结合进行研究,以达到电商企业的最新要求。电商企业每天都会产生海量的客户购买数据,通过分析客户购买行为来预测客户流失对于企业来说至关重要,数据挖掘技术在商业上的应用也由此而生。利用数据挖掘技术对电子商务网站的海量客户数据进行分析并研究,可以得出网购客户的流失预测模型,从而为电子商务服务商提供有价值的信息。数据挖掘技术是一种过程,这个过程整合了数学、统计、人工智能和机器学习的技术,从而可以在大型的数据库中提取和识别出对企业有用的信息。数据挖掘技术在客户关系管理上的应用已经成为了全球经济化时代的一个必然的趋势。数据挖掘技术是一种分析客户关系管理的有效工具,这种技术工具能够帮助企业储存和整合企业和客户之间的海量数据,分析出隐藏在这些海量数据下的大量信息,并帮助企业分析现有的客户,找出企业的潜在客户以及对企业有价值的高价值客户和浪费企业资源、却对企业没有任何盈利的负价值客户。这些信息的整合能够使得企业在高效的全球化进程中占据信息优势,更好地帮助企业提高企业资源的利用效率,提高企业营销政策的效果。本文基于数据挖掘技术构建客户流失预测模型,引入了客户关系管理理论中对客户历史购买行为进行描述的RFM理论,并结合当当网客户的实际情况,对网购客户的流失预测模型做出了修正,从而可以运用少数关键性指标对客户流失进行预测。本文的主要内容包括：(1)对已有的客户流失理论和技术进行总结;(2)对当当网客户购买数据进行探索性分析；(3)基于数据挖掘技术分析当当网购客户,构建网购客户流失预测模型。各章节主要内容如下：第一章是绪论部分。主要说明本文的研究背景、研究问题、研究内容以及研究的成果。第二章是理论基础部分。主要对客户关系管理理论、客户流失预测理论和客户细分理论进行了概括性地论述。第三章主要研究网购客户流失预测模型的构建,重点描述了在客户流失预测中应用广泛的逻辑斯蒂回归、决策树、神经网络三种技术的理论。第四章对当当网客户购买数据的实证分析。本文基于RFM理论提取出客户的购买行为数据,并对客户的行为做出了分析。第五章是模型构建部分。本文主要是在现有的研究基础上,针对B2C平台的客户提出了结合RFM理论和数据挖掘技术的网购客户流失预测模型,并对模型的结果进行了评估。第六章对论文的研究工作进行了总结,并提出了研究内容的局限性,指出今后的研究方向。
[Abstract]:In recent years, with the continuous improvement of productivity and the rapid development of information technology, the Internet has become an important strategic resource in today's society. With the advent of the Internet era, the business environment of enterprises has undergone tremendous changes, e-commerce platform can be constructed. The simplicity, speediness and convenience of e-commerce model has attracted great attention. A lot of customers have turned to this new market and joined the army of online shopping.
On the platform of E-commerce, it is far from enough for an enterprise to win the business battle only by attracting new customers and increasing market share. E-commerce enterprises must also do a good job in preventing the loss of customers, and solve the problems of "entering" and "exiting" customers, so as to achieve the goal of effective customer management for E-commerce enterprises. Customer churn research is more concerned about traditional enterprises, but less about the B2C platform or C2C platform in this new business environment. In this era of online shopping has become a way of life, traditional research is difficult to be used in the field of e-commerce. Therefore, this paper will traditional customer churn prediction model and electronics. The business mode is combined to study to meet the latest requirements of e-commerce enterprises.
E-commerce enterprises produce a large amount of customer purchase data every day. It is very important for enterprises to predict customer churn by analyzing customer purchase behavior, and the application of data mining technology in commerce comes into being. Purchase customer churn prediction model, so as to provide valuable information for e-commerce service providers.
Data mining technology is a process, which integrates mathematics, statistics, artificial intelligence and machine learning technology, so as to extract and identify useful information from large databases. The application of data mining technology in customer relationship management has become an inevitable trend in the era of global economy. Data mining technology is an effective tool to analyze customer relationship management. This technology tool can help enterprises store and integrate massive data between enterprises and customers, analyze a large amount of information hidden in these massive data, and help enterprises analyze existing customers, identify potential customers of enterprises and value to enterprises. The integration of these information can make the enterprise occupy the information superiority in the highly efficient globalization process, and help the enterprise to improve the utilization efficiency of enterprise resources and improve the effect of enterprise marketing policy.
Based on the data mining technology, this paper constructs the customer churn prediction model, introduces the RFM theory which describes the customer's historical purchasing behavior in the customer relationship management theory, and amends the customer churn prediction model according to the actual situation of Dangdang customers, so that a few key indicators can be used to predict the customer churn in. Forecast.
The main contents of this paper include: (1) summarizing the existing customer churn theory and technology; (2) exploratory analysis of Dangdang customer purchasing data; (3) building a customer churn prediction model based on data mining technology.
The first chapter is the introduction. It mainly explains the research background, research problems, research contents and research results.
The second chapter is the theoretical basis. It mainly discusses the customer relationship management theory, customer churn prediction theory and customer segmentation theory.
Chapter 3 mainly studies the construction of customer churn prediction model for online shopping. It mainly describes the theory of Logistic Regression, Decision Tree and Neural Network which are widely used in customer churn prediction.
The fourth chapter is the empirical analysis of Dangdang's customer purchase data. Based on RFM theory, this paper extracts the customer purchase behavior data and analyzes the customer behavior.
The fifth chapter is the construction of the model. Based on the existing research, this paper proposes a customer churn prediction model for online shopping combined with RFM theory and data mining technology for B2C platform customers, and evaluates the results of the model.
Chapter 6 summarizes the research work of the paper, puts forward the limitations of the research content, and points out the future research direction.
【学位授予单位】：西南财经大学
【学位级别】：硕士
【学位授予年份】：2013
【分类号】：F713.36

【参考文献】