基于因子分析和聚类的客户细分研究

Customers Segmentation Based on Factor Analysis and Cluster

摘要: 为了有针对性地进行客户关系维护，文章选取在线竞拍客户为研究对象，运用因子分析法把评价客户的十几项指标概括为5项更容易解释的评价指标，并在因子分析的基础上对客户进行综合评价和聚类；针对聚类数目设置的盲目性，提出运用聚类准则函数来衡量聚类数目的质量；通过对实际竞拍数据的计算分析，实验结果表明基于因子分析和聚类准则函数的客户聚类数目更为理想，聚类结果表明该方法有更好的客户竞拍行为解释能力和科学的客户细分能力。

Abstract: In order to individual customer relationship maintenance, this paper selects auction customers as the research object, using the method of factor analysis to summarize the eleven evaluation index as five evaluation indexes that are more easy to explain and comprehensive evaluation of customer and clustering are based on factor analysis. Aiming at the blindness set the number of clusters, the quality clustering criterion function to measure the number of clusters is put forward. Through calculation and analysis of the actual data, the experimental results show that the number of customer clusters is more ideal based on factor analysis and clustering criterion function. The clustering results has better explain ability and scientific customer segmentation ability on customer bidding behavior.