基于数据挖掘技术的客户流失预警模型  

Data Mining Technology based Customer Churn Alarm Models

作　　者：刘志超[1,2] 王雷[1,2] 谷垒[1] 冯湘[1] 

机构地区：[1]湖南大学软件学院,长沙410082 [2]中国联合通信有限公司湖南分公司,长沙410001

出　　处：《微计算机信息》2011年第2期176-177,175,共3页Control & Automation

摘　　要：结合客户细分的思想,提出了一种基于慢启动的频繁模式挖掘算法,并基于该算法提出了一种新的电信企业客户流失预警模型,湖南某大型电信企业基于该客户流失预警,在其一个地市分公司进行了客户维系与挽留二期工程的实施试点,试点结果表明,提出的客户流失预警模型具有良好的预警功能。By combing with the idea of customer Segmentation, a novel Slow-Start model based frequent pattern mining algorithm is proposed, and based on which, a new customer churn alarm model is proposed. According to the new customer churn alarm model, a major telecom in Hunan province, as an experiment area, has finished its second-stage project of customer retention. The results show that the newly proposed customer churn alarm model has good alarm performances.

关 键 词：客户流失 客户维系与挽留 预警模型 电信企业 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]