摘要
为研究用户用电行为数据挖掘问题,提出一种基于深度学习和改进蝗虫优化算法的用户电力数据挖掘方法。利用皮尔逊相关系数法选取用电行为特征集合,最大限度地降低数据处理维度。设计线性加权KFCM算法,采用改进蝗虫优化算法初始化聚类中心和聚类个数以提升聚类效果。利用改进的KFCM对用户用电行为进行数据挖掘,为电网企业网决策提供数据支撑。仿真实验结果表明,所提方法在聚类效果上有较好的表现。
The data mining of power consumption characteristics of massive users is studied,and a user power data mining method based on deep learning and improved grasshopper optimization algorithm is proposed.The Pearson correlation coefficient method is used to select the feature set of power consumption behavior to minimize the dimension of data processing.The linear weighted KFCM algorithm is designed,and the grasshopper optimization algorithm is used to initialize the cluster center and the number of clusters,so as to improve the clustering effect.The improved KFCM is used for data mining of users'power consumption behavior to provide data support for power grid enterprise network decision-making.Simulation results show that the proposed method has a good performance in clustering effect.
作者
王文
杨少杰
黄建平
WANG Wen;YANG Shaojie;HUANG Jianping(State Grid Zhejiang Electric Power Co.,Ltd.,Hangzhou 310007,Zhejiang)
出处
《微型电脑应用》
2024年第5期33-36,共4页
Microcomputer Applications
基金
浙江省自然科学基金项目(21ZJX1203)。
关键词
用电行为
聚类分析
蝗虫优化算法
特征选取
electricity consumption behavior
cluster analysis
grasshopper optimization algorithm
feature selection
