电信客户流失量估计组合模型的仿真研究  

Simulation Study on Combination model for Telecom Customer Churn Estimation

作　　者：王月敏[1,2] 

机构地区：[1]苏州大学计算机学院,江苏苏州215325 [2]江苏宿迁学院计算机科学系,江苏宿迁223800

出　　处：《计算机仿真》2013年第3期209-212,共4页Computer Simulation

摘　　要：研究电信客户流失问题,电信客户流失数据具有模糊性和非线性,单一算法仅能对模糊性或非线性进行预测,为提高电信客户流失估计准确率,提出了一种电信客户流失组合估计模型。首先对客户属性进行清洗并进行离散化处理,接着使用粗糙集方法对离散属性进行约简,刻画电信客户流失数据的模糊性;然后遗传算法优化支持向量机对电信客流失非线性进行描述,建立电信客户流失估计模型。仿真结果表明,粗糙集与支持向量机相融合模型克服单一粗糙集算法或支持向量机存在的缺陷,提高电信客户流失估计模型的估计准确率,可为电信客户管理优化设计提供依据。Aiming at the problems of the complexity of telecommunication customer loss and to improve prediction accuracy rate, this paper proposed a new telecom customer churn prediction model based on rough set theory and sup- port vector machine （SVM）. Firstly, customer properties were cleaned and discretely processed; then discrete attrib- utes were reduced by using rough set method, and the parameters of SVM were optimized by genetic algorithm to ob- tain the optimal parameters. Finally, the performance of model was tested. The experimental results show that the hit rate, covering rate, accuracy rate and lift coefficient of the proposed model are higher than those of the reference mod- el.

关 键 词：客户流失 粗糙集理论 支持向量机 遗传算法 

分 类 号：TM319[电气工程—电机]