Scholarlink

The optimization of resource utilization in residential buildings, which account for a sizable amount of the world’s energy consumption, is the main emphasis of this study. The research investigates the combination of Long Short-Term Memory (LSTM) neural networks and Federated Learning (FL) approaches. To improve prediction accuracy, LSTM neural networks can be used to better capture the intricate patterns in household electricity usage that are frequently missed by traditional methods. Moreover, FL facilitates cooperative model training across decentralized data sources, addressing privacy and data security concerns. By combining FL with LSTM neural networks, the paper seeks to provide a framework for precisely and effectively forecasting energy usage in residential contexts. Evaluations utilizing real-world datasets have proven the efficacy of this technique, opening the door for data-driven energy management solutions that put accuracy and privacy first.

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Predictive Analysis of Energy Consumption Using Long Short-Term Memory and Federated Learning

  • 会议论文

  • 发表日期：

    2025年

查看全文
  • Samhitha Nair,
  • Harry Thomas Kandathil,
  • P. Vijayakumar

摘要

The optimization of resource utilization in residential buildings, which account for a sizable amount of the world’s energy consumption, is the main emphasis of this study. The research investigates the combination of Long Short-Term Memory (LSTM) neural networks and Federated Learning (FL) approaches. To improve prediction accuracy, LSTM neural networks can be used to better capture the intricate patterns in household electricity usage that are frequently missed by traditional methods. Moreover, FL facilitates cooperative model training across decentralized data sources, addressing privacy and data security concerns. By combining FL with LSTM neural networks, the paper seeks to provide a framework for precisely and effectively forecasting energy usage in residential contexts. Evaluations utilizing real-world datasets have proven the efficacy of this technique, opening the door for data-driven energy management solutions that put accuracy and privacy first.