摘要
以某电信公司的历史数据为对象,运用数据挖掘技术,建立了基于决策树、神经网络和Logistic回归的三种客户流失预测模型.并对高价值高流失概率的客户进行K-means聚类分析,得到具有不同使用特征的五群客户,为对不同的流失客户群体提供针对性的营销策略提供了依据.
This paper uses data mining technique to analyze churn problems according to the historical data of a telecom corporation. By building different churn models based on decision tree, neural network and logistic regression, this paper is able to predict the potential customers. Finally, cluster models are adopted to deal with different customer＇ s groups in marketing, using k-means algorithm for customers who have high value and high churn probability according to their behaviors.
出处
《东莞理工学院学报》
2008年第1期77-81,共5页
Journal of Dongguan University of Technology
