摘要
运用结合PSO(粒子群优化)算法的模糊均值聚类法进行客户聚类分析是CRM中一个新的研究方向。本文提出将M个客户记录指定字段中出现频率最大的N个字段值作为客户的特征属性,由M个客户的特征属性构成客户模糊聚类的模式样品集,并在均值聚类算法中结合PSO算法,对总的类内离散度和进行优化,使其达到最小值,从而获取最佳客户聚类。实验表明,采用本算法能够得到满意的客户聚类结果。
Applying the fuzzy means clustering algorithms combined with PSO to the customer-clustering analysis in CRM is a new research field. This paper proposes an algorithm in which N keywords which appear most frequently inMcustomers are regarded as the features of the customers. The features of M customers compose a pattern sample set for fuzzy customer-clustering. The Particle Swarm Optimization algorithm is embedded into the fuzzy K-mean clustering algorithm so as to optimize the total scattering degree of clusters to be minimum and obtain the optimization of customer clusters. The resuits of experiments indicate that the algorithm can obtain better clustering results for the customer-clustering problems.
出处
《计算机工程与科学》
CSCD
北大核心
2009年第12期74-76,共3页
Computer Engineering & Science
基金
湖南省自然科学基金资助项目(07JJ3120)
湖南省科技计划资助项目(08GK3085)
湖南省教育厅资助项目(08C102)
