摘要
针对电信客户日益严重地流失问题,通过某电信运营商的历史资料,对电信PAS流失客户的自然属性和行为属性进行研究,利用决策树算法建立了客户流失预测模型。通过对模型进行评估分析,得到预测效果较好的模型,最后加入成本因素,进一步优化了模型。
According to the problem that the loss of customers in telecom is becoming more and more serious, this paper uses the history data in a special telecom provider＇s database, by analyzing the natural attributes and action attributes of the PAS lost cus-tomers, to establish the customer loss prediction models with decision tree algorithm. After the evaluation and analysis, it finds the better model, and with the considering of cost factor, the model is optimized in the end.
出处
《山西电子技术》
2009年第1期48-50,共3页
Shanxi Electronic Technology
关键词
数据挖掘
客户流失
决策树
预测模型
data mining
loss of customers
decision tree
prediction model
