Personalized Federated Learning for Heterogeneous Residential Load Forecasting

作     者：Xiaodong Qu Chengcheng Guan Gang Xie Zhiyi Tian Keshav Sood Chaoli Sun Lei Cui 

作者机构：the Shanxi Key Laboratory of Advanced Control and Equipment IntelligenceTaiyuan University of Science and TechnologyTaiyuan 030024China the Faculty of Engineering and Information TechnologyUniversity of Technology SydneyUltimo 2007Australia the Centre for Cyber Security Research and InnovationDeakin UniversityMelbourne 3125Australia the School of Computer Science and TechnologyTaiyuan University of Science and TechnologyTaiyuan 030024China 

出 版 物：《Big Data Mining and Analytics》 (大数据挖掘与分析（英文）)

年 卷 期：2023年第6卷第4期

页      面：421-432页

核心收录：

学科分类：12[管理学] 1201[管理学-管理科学与工程(可授管理学、工学学位)] 081104[工学-模式识别与智能系统] 08[工学] 0838[工学-公安技术] 0835[工学-软件工程] 0811[工学-控制科学与工程] 0812[工学-计算机科学与技术（可授工学、理学学位）] 

基　　金：supported by the Scientific and Technologial Innovation Programs of Higher Education Institutions in Shanxi,China(No.2020L0338) the Shanxi Key Research and Development Program(Nos.202102020101002 and 202102020101005) 

主　　题：load forecasting personalized federated learning differential privacy 

摘      要：Accurate load forecasting is critical for electricity production,transmission,and *** learning(DL)model has replaced other classical models as the most popular prediction ***,the deep prediction model requires users to provide a large amount of private electricity consumption data,which has potential privacy *** nodes can federally train a global model through aggregation using federated learning(FL).As a novel distributed machine learning(ML)technique,it only exchanges model parameters without sharing raw ***,existing forecasting methods based on FL still face challenges from data heterogeneity and privacy ***,we propose a user-level load forecasting system based on personalized federated learning(PFL)to address these *** obtained personalized model outperforms the global model on local ***,we introduce a novel differential privacy(DP)algorithm in the proposed system to provide an additional privacy *** on the principle of generative adversarial network(GAN),the algorithm achieves the balance between privacy and prediction accuracy throughout the *** perform simulation experiments on the real-world dataset and the experimental results show that the proposed system can comply with the requirement for accuracy and privacy in real load forecasting scenarios.