文档介绍：摘要
摘要
客户是企业最重要的资源之一。现代企业之间的竞争主要表现为
对客户的全面争夺,而是否拥有客户取决于企业与客户之间关系的状
况。企业要改善与客户之间的关系,就必须进行客户关系管理。客户
分析是客户关系管理的基础,而客户分析的一项重要内容是客户细
分,但目前还没有有效的客户细分方法。
本文在分析传统客户细分方法、客户忠诚度理论和数据挖掘算法
的基础上,探讨了以客户生命周期价值为依据的客户细分方法,并对
现有客户细分模型加以改进,提出了新的客户细分模型。新的客户细
分模型,依据客户的当前价值、潜在价值和忠诚度把客户分为八类,
进而针对不同类别客户提出了不同的市场营销策略。
对客户潜在价值的计算,我们采用关联规则算法,根据客户的历
史购买记录,预测出该客户将来可能购买的所有产品及购买概率,最
后再根据产品成本数据计算出客户的潜在价值。对于客户忠诚度的评
价,我们在国内外研究的基础上,提出了客户忠诚度评价的全新指标
体系,并根据这些指标数据分别运用了聚类、决策树和神经网络三种
数据挖掘算法对客户忠诚度进行预测和评价。
最后,结合某一个具体企业对客户细分模型进行了实例验证。
关键字:客户细分关联规则聚类决策树神经网络
I
华侨大学管理学硕士学位论文
Abstract
The customer is one of the most important resources of an enterprise.
petition between enterprises focuses on customers. Whether an
enterprise holds a customer depends on the state of the relationship
between the enterprise and the customer. To improve on the relationship
with the customer, an enterprise must carry out customer relationship
management (CRM).Customer analysis is the basis of CRM and customer
segmentation is an important item of customer analysis. But there is not
any efficient customer segmentation method at present.
The paper firstly introduces customer loyal theory, Data Mining
arithmetic and the traditional methods of customer segmentation, then
studies the customer-segmentation method according to customer lifetime
value and improves on the existing customer segmentation model. Based
on customer current value, customer potential value and customer loyal,
the new customer segmentation model divides customers into eight
classes, then according to the different customers’ class, the enterprise can
take different marketable strategy.
According to the customer's purchase history records, we use the
association rules algorithm to predict the future purchase and the
probability of purchase. Finally, we can get the customer potential value
according as product cost. Based on the present study status on the home
and abroad research, we