Privacy-preserving energy optimization via multi-stage federated learning for micro-moment recommendations
<p>Human behavior significantly impacts domestic energy consumption, making it essential to monitor and improve these consumption patterns. Traditional methods often rely on centralized servers to gather and analyze consumption data, which can lead to significant privacy risks as personalized...
Saved in:
| Main Author: | |
|---|---|
| Other Authors: | , , , |
| Published: |
2025
|
| Subjects: | |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| Summary: | <p>Human behavior significantly impacts domestic energy consumption, making it essential to monitor and improve these consumption patterns. Traditional methods often rely on centralized servers to gather and analyze consumption data, which can lead to significant privacy risks as personalized information becomes accessible online. To address this challenge, this study aims to optimize household energy consumption while preserving data privacy by proposing an innovative two-stage Federated Learning (FL) framework that delivers real-time micro-moment-based recommendations. Leveraging FL enables efficient model training across diverse end-user applications while preserving data privacy. The proposed framework employs a two-stage FL training methodology, utilizing the DRED and QUD datasets, and achieves substantial performance improvements. A comparative evaluation of three FL algorithms (FedAvg, FedProx, Mime-lite) identifies the most suitable aggregation strategy. The model achieves robust performance, with approximately 98 % accuracy and F1-score in the second training stage. These findings demonstrate the effectiveness of FL in enabling personalized, privacy-preserving energy recommendations. The novelty of this work lies in combining micro-moment prediction with a multi-stage FL architecture tailored for smart home energy optimization. This study highlights the potential of FL to enhance energy efficiency and sustainability while safeguarding user privacy, paving the way for future research in energy optimization and sustainable living.</p><h2>Other Information</h2> <p> Published in: Sustainable Energy, Grids and Networks<br> License: <a href="http://creativecommons.org/licenses/by/4.0/" target="_blank">http://creativecommons.org/licenses/by/4.0/</a><br>See article on publisher's website: <a href="https://dx.doi.org/10.1016/j.segan.2025.102100" target="_blank">https://dx.doi.org/10.1016/j.segan.2025.102100</a></p> |
|---|