利用大数据技术提升电力客户档案资源管理和服务能力  

Using Big Data Technology to Improve Power Customer File Resource Management and Service Capabilities

作　　者：陈明 刘睿 李乐 李锐锋 曾琴 李玉婷 CHEN Ming;LIU Rui;LI Le;LI Ruifeng;ZENG Qin;LI Yuting(Jiuquan Power Supply Company of State Grid Gansu Electric Power Company,Jiuquan 735000,Gansu,China;Chengdu Kepuwei Information Technology Co.,Ltd.,Chengdu 610042,Sichuan,China)

机构地区：[1]国网甘肃省电力公司酒泉供电公司,甘肃酒泉735000 [2]成都科普威信息技术有限责任公司,四川成都610042

出　　处：《电力大数据》2022年第2期9-18,共10页Power Systems and Big Data

摘　　要：为解决营销业务人员填报流程信息错误、快速查出营销系统客户档案不准确信息,减少人为原因填报不准确造成的电价执行错误及电费纠纷问题,提升客户档案数据的可用性,基于营销系统、用电采集系统档案数据,对数据清理、集成、规约、离散化完成数据清洗和规范化转换,形成电力客户档案数据字段宽表,并以数据结构、关联规则、主从关系为主线,梳理不同字段的单相关、偏相关、复相关关系和数据特性,探索字段之间的关联关系。通过AdaBoost分类器、Knuth-Morris-Pratt算法、IF-THEN规则、Sunday算法等多种大数据技术,对整理的数据进行分析、统计并构建完整的数据应用模型,形成业务规则模型及数据特性分析模型,实现自动统计分析异常历史数据可视化展示,提升档案异动数据整改效率;业务智能分析模型实现营销业务流程输入字段自动输出相关联字段信息,提高填报准确率。This paper is dovted to solve the problem of incorrect information in the process information reported by marketing personnel,quickly find inaccurate information in the customer files of the marketing system,reduce the implementation of electricity prices and electricity tariff disputes caused by inaccurate reporting of human reasons,and improve the availability of customer’s file data.Based on the archive data of the marketing system and the electricity consumption data aquisition system,the clear and standardized conversion of data cleaning,integration,specification,and discretization is completed to form a wide table of data fields in the power customer archives.To take the the data structure,association rules,and master-slave relationship as the main line,we can sort out the single correlation,partial correlation,multiple correlation relationship and data characteristics of different fields,and explore the relationship between fields.Through AdaBoost classifier,Knuth-Morris-Pratt algorithm,IF-THEN rule,Sunday algorithm and other big data technologies,analyze and count the sorted data,we can build a complete data application model,business rule model and data characteristic analysis model.The automatic statistical analysis of the visual display of abnormal historical data can be realized,improving the efficiency of file change data rectification.By business intelligence analysis model,the automatic output of related field information in the input fields of the marketing business process also can be realizes,improving the accuracy of reporting.

关 键 词：客户档案 大数据 相关性 分类器 决策树 可视化 

分 类 号：TM714[电气工程—电力系统及自动化]