决策树算法与客户流失分析  

Decision tree algorithm and churn analysis

作　　者：邓全 

机构地区：[1]中国电信陕西公司网运部,陕西西安710065

出　　处：《西安邮电学院学报》2013年第3期49-51,共3页Journal of Xi'an Institute of Posts and Telecommunications

摘　　要：为了向市场人员提供决策依据,有效降低客户流失率,基于数据挖掘平台Clementine,构造出一种客户流失模型,并分别利用决策树算法C5.0及分类和回归算法对某运营商提供的实际数据进行实验分析。对比实验结果可知,C5.0算法在准确率及覆盖率等方面更适合于该运营商。In order to provide the market basis for decision making and to reduce customer churn, a kind of customer churn model is constructed based on data mining platform Clementine. The ac- tual data provided by some operators are analysed and calculated by using the C5.0 algorithm and the classification and regression algorithm separately. The experimental results show that the C5.0 algorithm is more suitable to the operators in the aspects of accuracy and coverage.

关 键 词：数据挖掘 CLEMENTINE C5 0算法 分类和回归树 决策树 

分 类 号：TP391.13[自动化与计算机技术—计算机应用技术]