The IRMA Community
Newsletters
Research IRM
Click a keyword to search titles using our InfoSci-OnDemand powered search:
|
Residential Electricity Consumption Prediction Method Based on Deep Learning and Federated Learning Under Cloud Edge Collaboration Architecture
|
|
Author(s): Wei Wang (State Grid Hebei Marketing Service Center, China), Xiaotian Wang (State Grid Hebei Marketing Service Center, China), Xiaotian Ma (State Grid Hebei Marketing Service Center, China), Ruifeng Zhao (State Grid Hebei Marketing Service Center, China)and Heng Yang (Beijing Tsingsoft Technology Co., Ltd., Beijing, China)
Copyright: 2024
Volume: 16
Issue: 1
Pages: 19
Source title:
International Journal of Gaming and Computer-Mediated Simulations (IJGCMS)
Editor(s)-in-Chief: Hui Li (Beijing University of Chemical Technology, China)
DOI: 10.4018/IJGCMS.336846
Purchase
|
Abstract
Traditional residential electricity prediction methods have problems, such as difficulty in ensuring user privacy and poor convergence speed due to the influence of Different Residential Electricity Consumption (REC) habits. A REC prediction method based on Deep Learning (D-L) and Fed-L under the Cloud Edge Collaboration (CEC) architecture is proposed to address the above issues. First, based on the CEC architecture, combining edge computing and cloud computing center server, the overall model of REC prediction is built. Then, Federated Learning (Fed-L) and D-L model Empirical Mode Decomposition - Long Short-Term Memory (EMD-LSTM) were introduced on the edge side, and the edge side Fed-L depth model was personalized by using EMD-LSTM. Finally, aggregation of edge side models was achieved in the cloud by receiving encrypted model parameters from the edge side and updating and optimizing all edge side models. The results show that the proposed method has the highest REC prediction accuracy, reaching 96.56%, and its performance is superior to the other three comparative algorithms.
Related Content
|
Zhenjiang Cao, Zhenhai Cao.
© 2024.
17 pages.
|
|
Wei Guo, Shengbo Sun, Peng Tao, Fei Li, Jianyong Ding, Hongbo Li.
© 2024.
19 pages.
|
|
Zheqi Zhu, Kun Zhai.
© 2024.
24 pages.
|
|
Wei Wang, Xiaotian Wang, Xiaotian Ma, Ruifeng Zhao, Heng Yang.
© 2024.
19 pages.
|
|
Poe Sriwatanathamma, Veerawat Sirivesmas, Sone Simatrang, Nobonita Himani Bhowmik.
© 2024.
18 pages.
|
|
Qiang Wang, Zhenwei Huang.
© 2024.
16 pages.
|
|
.
© 2024.
|
|
|