Hybrid Federated Learning Aggregation Methods for Non-IID Time Series Energy Demand Forecasting Using Gated Recurrent Unit (GRU) Neural Networks

Kumari, Shilpa (2025) Hybrid Federated Learning Aggregation Methods for Non-IID Time Series Energy Demand Forecasting Using Gated Recurrent Unit (GRU) Neural Networks. Masters thesis, Dublin, National College of Ireland.

Preview	PDF (Master of Science) Download (1MB) | Preview
Preview	PDF (Configuration Manual) Download (854kB) | Preview

Abstract

With the increasing deployment of smart meters and IoT infrastructure, time series data on energy consumption is being generated at a massive scale. While centralized machine learning models are traditionally used for demand forecasting, privacy concerns, regulatory constraints, and the distributed nature of data necessitate a shift towards decentralized methods. Federated Learning (FL) addresses these concerns by allowing clients to collaboratively train models without sharing raw data. However, standard FL techniques such as FedAvg struggle when applied to non-IID (non-independent and identically distributed) time series data, leading to degraded model performance and slow convergence. This research proposes a hybrid framework that combines the FedProx federated optimization algorithm with Gated Recurrent Unit (GRU) neural networks for energy demand forecasting in non-IID environments. By using FedProx's proximal regularization during local training and GRU's capability to model temporal dependencies, the approach aims to enhance forecasting accuracy, reduce client model divergence, and achieve better generalization across diverse households or regions. Publicly available datasets from the Open Power System Data (OPSD) project will be used to simulate realistic federated settings.

Item Type:	Thesis (Masters)
Supervisors:	Name Email Subhnil, Shubham UNSPECIFIED
Subjects:	H Social Sciences > HC Economic History and Conditions > Natural resources > Power resources > Energy consumption T Technology > TK Electrical engineering. Electronics. Nuclear engineering > Telecommunications > Computer networks > Internet of things Q Science > Q Science (General) > Self-organizing systems. Conscious automata > Machine learning
Divisions:	School of Computing > Master of Science in Data Analytics
Depositing User:	Ciara O'Brien
Date Deposited:	25 Aug 2026 15:43
Last Modified:	25 Aug 2026 15:43
URI:	https://norma.ncirl.ie/id/eprint/9640

Actions (login required)

View Item