太赫兹科学与电子信息学报2025，Vol.23Issue(7)：692-698,7.DOI:10.11805/TKYDA2024595
基于多维能耗数据的用能行为聚类分析及数据降维方法
Cluster analysis of energy use behavior based on multidimensional energy consumption data and data dimensionality reduction methods
摘要
Abstract
Long-term energy consumption data recording and analysis help to identify the trends and patterns of energy consumption,providing important references for formulating carbon reduction strategies.The energy consumption data of key carbon emission monitoring users involves various types.It is not only necessary to conduct clustering analysis on the users'energy consumption data,but also to study more precise clustering result visualization methods.To this end,a Fuzzy C-Means algorithm based on Tent Chaotic Sequence Grey Wolf Optimization(TGWO-FCM)is proposed to analyze the users'energy consumption data.The clustering centers of users'energy consumption data are regarded as grey wolf individuals for optimization,which solves the shortcomings of the FCM algorithm being sensitive to the initial clustering center positions and easily falling into local optimum.The data dimensionality reduction method of Uniform Manifold Approximation Projection(UMAP)is adopted to reduce the complexity of energy consumption data,mapping high-dimensional energy consumption data to two-dimensional or three-dimensional spaces to achieve intuitive visualization of the data.Experimental results show that the method proposed in this paper can classify users with similar energy consumption patterns into the same category,not only revealing the differences in energy consumption patterns among users,but also providing a scientific basis for formulating targeted energy-saving and emission-reduction policies.关键词
混沌序列/灰狼算法/聚类/多维能耗数据/数据降维Key words
chaotic sequences/grey wolf algorithm/clustering/multidimensional energy data/data downscaling分类
信息技术与安全科学引用本文复制引用
罗帅,王洋,项添春,周进,李娜,张来..基于多维能耗数据的用能行为聚类分析及数据降维方法[J].太赫兹科学与电子信息学报,2025,23(7):692-698,7.基金项目
国网天津市电力公司科技项目资助(经研-研发2024-02) （经研-研发2024-02）
