基于客户流失分析的聚类分析模型的构建

A Clustering Mode of Missed Customer

本文针对电信公司的客户流失主题,采用了聚类方法对特征属性进行分类。鉴于聚类算法的无监督性,结合因子方差分析方法对聚类进行监督,分析特征属性与目标变量的相关性,利用相关属性与目标变量共同参与聚类算法构建聚类模型,获取流失客户的特征。

Aimed at the customer missing in telecom, clustering is used to classify the characterized vector in this article, and for the non- supervision of clustering, factor variance analysis is used to analyze the relevance between characterized vector and target variable, and then the relevant characterized vector and target variable are used together to establish the clustering model, and finally the features of missed customer were acquired.