基于二值属性的聚类分析算法

Clustering algorithm based on two-value property

解决分析型CRM系统中客户类别划分问题,聚类分析是必不可少的算法,根据分析对象所涉及的数据类型、聚类目的以及具体应用要求选择恰当的聚类算法是解决问题的关键.针对客户关系管理中依据客户购买力划分客户类别的问题,结合相关营销理论、数据类型及属性表示方法,提出使用二值属性描述对象、采用最短距离法进行聚类的算法,并通过该方法进行了实例计算.计算结果表明,用二值属性表示对象的方法不仅简洁、方便、易懂,而且对聚类结果具有直观的检验性.

Clustering analysis is a necessary algorithm to solve distinguishing customers in an analytical customer relationship management (CRM) system. It is critical to select a suitable algorithm according to data type, clustering goal, and applied demand. In view of distinguishing customers based on buying abilities in a CRM system, an algorithm was put forward that a two-value property was used to describe objects and then the min-distance algorithm was used to cluster according to selling principles, data-type expression and properties expression. An example was given to examine this algorithm. Calculations show that this method is simple, convenient, and easy to understand, but also make it possible to examine clustering results directly.