摘要
目前,电力用户的缴费方式仍以线下为主,长期以来造成了电力企业运营成本高、管理压力大等问题。传统的线上引流方法用户增长速度慢,用户粘性差。为快速转变电力用户长期以来所形成的线下缴费方式,提出采用基于群簇核心用户的推荐模型。电力用户的缴费行为表现出一定的群聚性和周期性,群簇核心用户不仅可以精准地代表某一类簇的行为特征,还可以间接地影响相关用户,实现新型缴费方式的快速传播,从而加速线上用户的增长速度。此外,提出的模型具有速度快、可并行、易迁移、易操作等特点,实现了一户一方案的精准推荐,在大规模电力用户数据下表现出了极高的应用价值。仿真实验结果表明,所提算法在进行电力用户精准引流时,在召回率和精确度等指标上具有较好的效果。
The payment mode of power users is mainly offline leads to the high operation cost and management pressure of power enterprises.Traditional online drainage methods have slow user growth and poor user stickiness.To quickly transform the offline payment mode formed by power users for a long time,the recommendation model based on cluster kernel users was proposed.The payment behavior of power users showed certain clustering and periodicity.Kernel users of the cluster could not only accurately represent the behavior characteristics of a certain type of cluster,but also indirectly influence relevant users to realize the rapid spread of the new payment method,thus the growth rate of online users was accelerated.In addition,the proposed model had the characteristics such as fast speed,parallelism,easy migration and easy operation,which could realize the accurate recommendation of one scheme for one household and show the extremely high application value under the data of large-scale power users.The simulation results indicated that the proposed algorithm had good effects on the recall rate and accuracy when carrying out accurate drainage of power users.
作者
宫立华
盛妍
李磊
刘鲲鹏
朱银龙
何薇
徐倩丽
GONG Lihua;SHENG Yan;LI Lei;LIU Kunpeng;ZHU Yinlong;HE Wei;XU Qianli(Customer Service Center, State Grid Corporation of China, Tianjin 300322, China;Department of Sales, State Grid Corporation of China, Beijing 100031, China;Sales Service Center, State Grid Shanxi Electric Power Company, Taiyuan 030009, China;China Power Puhua Information Technology Co., Ltd., Beijing 100085, China;Beijing Shuyang Intelligent Technology Co., Ltd., Beijing 100044, China)
出处
《计算机集成制造系统》
EI
CSCD
北大核心
2021年第9期2721-2728,共8页
Computer Integrated Manufacturing Systems
关键词
群簇核心用户
电力用户
推荐模型
精准引流
cluster kernel users
power users
recommendation model
accurate drainage