基于近邻传播聚类-K均值聚类的工业用户用电模式挖掘方法
Methods for Mining the Electricity Consumption Mode of Industrial Users Based on Affinity Propagation Clustering-K-Means Clustering
宗一 1郑罡 1南钰1
作者信息
- 1. 国网河南省电力公司开封供电公司 河南开封 471000
- 折叠
摘要
为了充分发挥用户负荷的可调节潜力,提出了一种基于近邻传播聚类-K均值聚类的工业用户用电模式挖掘方法.首先,比较K均值聚类和近邻传播聚类-K均值聚类的优缺点.在工业用户的选取上,选择最佳聚类数均为3的工业用户负荷数据作为被分析对象以便聚类,借助MATLAB工具对用户负荷数据进行聚类,得到了3组所需的聚类中心,再绘制成曲线以便观察和后续提取特征指标.
Abstract
This paper proposes a method of mining the electricity consumption mode of industrial users based on af-finity propagation clustering-K-means clustering to give full play to the adjustable potential of user loads.Firstly,it compares the advantages and disadvantages of k-means clustering and affinity propagation clustering-K-means clustering.In the selection of industrial users,it selects the industrial user load data with the best clustering number of 3 as the analyed object for clustering,clusters the user load data with the help of the MATLAB tool,and obtains the three groups of required cluster centers,which are then drawn into curves for observation and the subsequent extraction of characteristic indicators.
关键词
近邻传播聚类-K均值聚类/工业用户/可调节潜力评估/评估指标体系/多准则决策法Key words
Affinity propagation clustering-K-means clustering/Industrial user/Adjustable potential assessment/Evaluation index system/Multi-criteria decision method引用本文复制引用
基金项目
国家河南省电力公司科技项目(521790230003)
出版年
2024
NETL
NSTL
万方数据