基于数据挖掘的用户行为特征提取系统构建  

Construction of User Behavior Feature Extraction System Based on Data Mining

作　　者：陈景琪 丁凌 CHEN Jing-qi;DING Ling(State Grid Shanghai Electric Power Company,Shanghai 200120 China)

机构地区：[1]国网上海市电力公司,上海200120

出　　处：《自动化技术与应用》2021年第12期50-53,61,共5页Techniques of Automation and Applications

摘　　要：为充分缓解电量负荷压力,本文设计了基于数据挖掘的用户行为特征提取系统。按照电力信息提取框架的处置需求,连接用户行为管理模块与电力数据挖掘驱动单元,完成用户行为特征提取系统的硬件环境搭建。采用关联电力用户行为特征树,存储各类待挖掘的电力数据,完成系统的软件环境搭建,结合相关硬件设备结构,完成基于数据挖掘的用户行为特征提取系统设计。实验结果表明,与基于k-means的提取系统相比,应用新型特征提取系统后,电力用户的总负荷水平明显下降,电子转存频率也由47%提升至98%,用户端主体的电量负荷压力得到良好的平均与协调。In order to fully alleviate the pressure of electricity load, a user behavior feature extraction system based on data mining is designed. According to the disposal requirements of the power information extraction framework, the user behavior management module and the power data mining driving unit are connected to complete the hardware environment of the user behavior feature extraction system. Using the related power user behavior feature tree to store all kinds of power data to be mined,the software environment of the system is built, and the design of user behavior feature extraction system based on data mining is completed combining with the related hardware structure. The experimental results show that, compared with the K-means based extraction system, the total load level of the power users is significantly reduced, the electronic transfer frequency is also increased from 47% to 98%, and the power load pressure of the users is well averaged and coordinated.

关 键 词：数据挖掘 特征提取 行为特征树 数据存储 

分 类 号：TP311.13[自动化与计算机技术—计算机软件与理论] TP393.092[自动化与计算机技术—计算机科学与技术]