基于聚类实现客户行为分析  

Realizing customer behavior analysis based on clustering

作　　者：曲昭伟[1] 郑岩[2] 吕廷杰[1] 

机构地区：[1]北京邮电大学电子商务研究中心,北京100876 [2]北京邮电大学计算机科学与技术学院,北京100876

出　　处：《东北师大学报（自然科学版）》2006年第2期19-21,共3页Journal of Northeast Normal University(Natural Science Edition)

基　　金：国家自然科学基金资助项目(60175024);教育部重点实验室开放课题资助项目(F0607-01)

摘　　要：在介绍客户行为分析一般过程的基础上,提出了一种实现客户行为分析的聚类算法.该算法通过构造模糊相异矩阵,将高维、海量的客户消费模式映射到二维平面,进一步利用遗传算法(GA)进行优化,快速而灵活地实现客户行为的聚类分析,发现大客户.Customer behavior analysis is one of the most popular subjects in Customer Relationship Management （CRM）. A clustering algorithm of realizing the customer behavior analysis is presented based on introducing the general process of customer behavior analysis. In the presented clustering algorithm, the fuzzy dissimilarity matrix is constructed and the high dimensional and large amount consumption modes of the customer are mapped to the two - dimensional plane. Furthermore, the clustering algorithm is optimized by genetic algorithm （GA） in order to implement the customer behavior analysis and find the master customer flexibly and quickly.

关 键 词：遗传算法 模糊相异矩阵 客户行为分析 客户关系管理 

分 类 号：TP183[自动化与计算机技术—控制理论与控制工程]