摘要
针对企业CRM中日常收集的各种销售数据,设计一种反映客户价值与客户关系质量的客户细分模型。运用概念格获取最大频繁项目集,并以此作为初始聚类,采用适合的相似性测量方法,求得聚类结果。目的在于探讨基于概念格技术的客户聚类方法的可行性和有效性。研究表明,该方法所生成的聚类比其他传统方法更优化,而且效率较高。
Aiming at data of some enterprises dealing with whole sale business, designes a customer segmentation model which can reflect the value and quality of customer relations. Maximum frequent itemsets can be obtained via concept lattice. As the initial cluster, these maximum frequent itemsets can be used to acquire the clustering results with the similarity measurement method. Aiming at discussing the feasibility and effectiveness of customer segmentation provided by concept lattice. Study shows that the method not only generates optimal clusters than traditional algorithms, but also exhibits good efficiency.
出处
《现代计算机》
2008年第6期70-73,共4页
Modern Computer
关键词
概念格
最大频繁集
聚类
客户细分
Concept Lattice
Maximum Frequent Itemsets
Clustering
Customer Segmentation
