基于潜在类别模型的电信用户分类研究

Research on Telecom User Classification Based on Latent Class Model

电信用户的消费行为与消费特征不尽相同,如何对电信用户群体进行有效划分,是值得关注的问题。文章研究了利用电信用户的自身属性与业务属性,构建潜在类别模型,通过10个外显变量对其进行分类,并根据划分结果分析每一类用户的消费特征。研究结果表明,电信用户可分为临时型、低值型、传统型、公务型、佛系型和稳定型6种类型,其中公务型用户群体消费水平最高且相对稳定,具有较高的用户维护价值,运营商应给予最多的关注。

The consumption behavior and consumption characteristics of telecom users are different.How to divide telecom user groups effectively is a problem worthy of attention.Based on the self attributes of telecom users and service attribute,this paper constructs latent class model,classifies them by 10 explicit variables,and analyzes the consumption characteristics of each category of users according to the classification results.The results show that telecom users can be divided into six types:temporary,low value,traditional,official,buddhist and stable.Among them,the consumption level of official users is the highest and relatively stable,which has higher user maintenance value,and operators should pay more attention to it.