一种带有导向性的聚类方法在电信客户细分中的应用  

An application of a new oriented algorithm of clustering in telecom customer segmentation

作　　者：黄海[1] 林齐宁[1] 

机构地区：[1]北京邮电大学经济管理学院,北京100876

出　　处：《北京邮电大学学报（社会科学版）》2006年第1期48-51,共4页Journal of Beijing University of Posts and Telecommunications(Social Sciences Edition)

摘　　要：聚类是一种常用的客户细分方法。本文结合电信市场客户细分实践,针对一般聚类算法的不足,设计了一种带有导向性的聚类方法。使用该方法得出的聚类结果区分了普通类与极端类,提高了聚类效果,在指导实际市场营销活动中有着明显的优势。Clustering is a good method which has been widely used in customer segmentation. The paper puts forward a new oriented algorithm of clustering according to customer segmentation of telecom markets. The oriented algorithm of clustering may distinguish the normal clusters and the uttermost clusters. Using this oriented algorithm, the clustered result can be much more scientific and intelligible, which has been proved by some experimentation.

关 键 词：客户细分 聚类 导向性 普通类 极端类 电信市场 

分 类 号：F626[经济管理—产业经济]