基于数据挖掘技术的B2C企业客户关系管理研究

发布时间：2018-07-05 18:57

本文选题：数据挖掘 + 客户分类　；参考：《沈阳工业大学》2016年硕士论文

【摘要】：随着电子商务的迅猛发展和信息处理技术能力的增强,数据挖掘技术在商务领域得到了广泛的应用,这一应用能够提升商家对客户的识别、分析和需求满足能力。但是由于不同行业的数据采集点、数据处理要求及结果体现形式的不同,需要针对不同的行业甚至是企业构建适合自身的模型和运算方法。本文总体上在分析B2C客户消费特点的基础上,以电子商务相关理论为指导,综合运用几种数据挖掘技术于企业客户关系管理中,以期望技术的运用提升企业的客户关系管理水平。本文在介绍相关概念和技术手段的基础上,首先分析了B2C下客户消费的特点,并着重分析了客户分析的流程,分别分析了客户分析的总体流程和RFM理论下的客户数据分析流程,并基于RFM视角下客户价值和基于客户属性、消费心理特征和网络影响等非价值指标对客户进行了分类,并构建了基于二维聚类的客户分类模型。其次,基于不同分类,针对于客户维护重点是要利用关联规则对客户的潜在需求进行挖掘和分析,并试探性的从扫描次数减少的角度对关联规则的Apriori算法进行了改进；针对新客户主要是根据其基本注册信息和浏览记录利用个性化推荐技术进行推荐,并考虑数据量的影响引入规模因子改进了相关系数的计算,以提升预测的精度。再次,利用分类结果和关联规则构建了退出类、重点类、普通类、潜力类及黄金类五类客户关系管理的策略,并提出了实施的保障。最后通过案例分析证明了本文提出方法的有效性。本文试图采用双聚类组合的方式对客户进行分类,并探索改进关联规则和推荐系统的算法,为B2C企业的客户关系管理水平提升做出一定的支撑。
[Abstract]:With the rapid development of electronic commerce and the enhancement of information processing technology, data mining technology has been widely used in the field of commerce. This application can improve the ability of merchants to identify, analyze and meet the needs of customers. However, due to the difference of data collection points, data processing requirements and the results of different industries, it is necessary to build suitable models and operation methods for different industries or even enterprises. Based on the analysis of the characteristics of B2C customer consumption, and guided by the related theory of electronic commerce, this paper synthetically applies several kinds of data mining techniques to enterprise customer relationship management. Improve the level of customer relationship management with the application of expected technology. Based on the introduction of related concepts and technical means, this paper first analyzes the characteristics of customer consumption under B2C, and focuses on the process of customer analysis, and analyzes the overall flow of customer analysis and the flow of customer data analysis based on RFM theory, respectively. Customers are classified based on customer value and non-value indicators such as customer attributes, consumer psychological characteristics and network influence from the perspective of RFM, and a customer classification model based on two-dimensional clustering is constructed. Secondly, based on different classification, the focus of customer maintenance is to mine and analyze the potential needs of customers by using association rules, and tentatively improve the Apriori algorithm of association rules from the angle of reducing scanning times. According to the basic registration information and browsing record, the new customers are recommended by personalized recommendation technology, and the scale factor is introduced to improve the calculation of correlation coefficient to improve the accuracy of prediction. Thirdly, by using classification results and association rules, the strategies of customer relationship management of exit class, key class, common class, potential class and gold class are constructed, and the security of implementation is put forward. Finally, the effectiveness of the proposed method is proved by a case study. In this paper, we try to classify customers by using double clustering and combination, and explore the algorithm of improving association rules and recommendation system to support the improvement of customer relationship management level in B2C enterprises.
【学位授予单位】：沈阳工业大学
【学位级别】：硕士
【学位授予年份】：2016
【分类号】：TP311.13;F274

【参考文献】