Shizhen Hu, Guangdi Li, Haoyi Wang, Hongyuan Ma, Ziwen Li
{"title":"Demand response intelligence recommendation based on knowledge graph and knowledge graph convolutional neural network","authors":"Shizhen Hu, Guangdi Li, Haoyi Wang, Hongyuan Ma, Ziwen Li","doi":"10.1109/ICCSIE55183.2023.10175209","DOIUrl":null,"url":null,"abstract":"A demand response intelligent recommendation model integrating knowledge graph and knowledge graph neural network(KGCN) is proposed to address the problems of cold start and sparsity of existing demand response intelligent recommendation algorithms. The structured triad of users’ electricity consumption is extracted from the users’ electricity consumption data set, followed by clustering users with similar electricity consumption behaviors through an improved clustering algorithm, and adding the clustering results to the knowledge graph together with the structured triad, using the KGCN model to embed the neighborhood entity information into the vector space to solve the data sparsity problem; meanwhile, using the prior knowledge in the graph to solve the cold start problem; to solve the To solve the recommendation lag problem, multi-hop propagation algorithm is introduced to reduce the set of candidate users and improve the recommendation efficiency. The results show that the intelligent recommendation model based on KGCN and knowledge graph can effectively solve the above problems and improve the indexes compared with the existing traditional algorithms.","PeriodicalId":391372,"journal":{"name":"2022 First International Conference on Cyber-Energy Systems and Intelligent Energy (ICCSIE)","volume":"42 1","pages":"0"},"PeriodicalIF":0.0000,"publicationDate":"2023-01-14","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":"0","resultStr":null,"platform":"Semanticscholar","paperid":null,"PeriodicalName":"2022 First International Conference on Cyber-Energy Systems and Intelligent Energy (ICCSIE)","FirstCategoryId":"1085","ListUrlMain":"https://doi.org/10.1109/ICCSIE55183.2023.10175209","RegionNum":0,"RegionCategory":null,"ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"","JCRName":"","Score":null,"Total":0}
引用次数: 0
Abstract
A demand response intelligent recommendation model integrating knowledge graph and knowledge graph neural network(KGCN) is proposed to address the problems of cold start and sparsity of existing demand response intelligent recommendation algorithms. The structured triad of users’ electricity consumption is extracted from the users’ electricity consumption data set, followed by clustering users with similar electricity consumption behaviors through an improved clustering algorithm, and adding the clustering results to the knowledge graph together with the structured triad, using the KGCN model to embed the neighborhood entity information into the vector space to solve the data sparsity problem; meanwhile, using the prior knowledge in the graph to solve the cold start problem; to solve the To solve the recommendation lag problem, multi-hop propagation algorithm is introduced to reduce the set of candidate users and improve the recommendation efficiency. The results show that the intelligent recommendation model based on KGCN and knowledge graph can effectively solve the above problems and improve the indexes compared with the existing traditional algorithms.