采用改进高斯混合模型的电力客户大数据行为分析  

Big data behavior analysis of power customers based on improved gaussian mixture model

作　　者：吉涛 刘玮洁 段立 郑伟 廖勇[3] JI Tao;LIU Weijie;DUAN Li;ZHENG Wei;LIAO Yong(State Grid Chongqing Electric Power Company Information&Telecommunication Branch,Chongqing 401120,China;Tongliang Power Supply Branch of State Grid Chongqing Electric Power Company,Chongqing 402560,China;School of Microelectronics and Communication Engineering,Chongqing University,Chongqing 400044,China)

机构地区：[1]国网重庆市电力公司信息通信分公司,重庆401120 [2]国网重庆市电力公司铜梁供电公司,重庆402560 [3]重庆大学微电子与通信工程学院,重庆400044

出　　处：《重庆理工大学学报（自然科学）》2022年第5期233-240,共8页Journal of Chongqing University of Technology：Natural Science

基　　金：国网重庆市电力公司科技项目(2021渝电科技8#)。

摘　　要：随着电网系统的不断完善及用户数的不断增加,智能电网系统中存储的客户信息逐渐形成客户大数据,从这些数据中可以分析得到用户用电行为等一些潜在信息,因此如何从中挖掘出这些隐藏信息并利用此类信息来提升公司的效率成为本文研究重点。提出一种联合基于密度的带噪空间聚类(density-based spatial clustering of application with noise,DBSCAN)算法与期望最大化(expectation maximization,EM)算法的高斯混合聚类算法,通过DBSCAN算法确定合适的k个聚类中心及迭代初始数据,再通过EM算法迭代出聚类结果。案例分析表明:和其他几种典型聚类算法相比,所提算法在分析大数据和挖掘电力客户用电行为信息方面更加快速和准确,可以更有效地对电力公司客户行为数据进行聚类分析。With the improvement of the power grid system and the increase of users,the customer information stored in the smart grid system has gradually formed customer big data.From this data,we can analyze and obtain some potential information such as users’electricity consumption behavior.Therefore,how to dig out these hidden information and use such information to improve the efficiency of the company has become the research focus of this paper.This paper proposes a Gaussian hybrid clustering algorithm that combines the density-based spatial clustering of application with noise(DBSCAN)algorithm and expectation-maximization(EM)algorithm.The suitable k clustering centers and initial data of iteration are determined by the DBSCAN algorithm.Then,the clustering results are obtained by the EM algorithm.The case study shows that,compared with other typical clustering algorithms,the proposed algorithm can quickly and effectively analyze the big data of power users,and mine the effective information of users’electricity consumption behavior,which can more effectively cluster the customer behavior data of power companies.

关 键 词：电力系统 大数据 客户行为分析 高斯混和聚类 

分 类 号：TM73[电气工程—电力系统及自动化]