본문 바로가기
  • Home

A Multi-Head RNN-based Architecture for Short-term Electricity Demand Forecasting

  • Journal of The Korea Society of Computer and Information
  • Abbr : JKSCI
  • 2025, 30(12), pp.139~146
  • Publisher : The Korean Society Of Computer And Information
  • Research Area : Engineering > Computer Science
  • Received : October 17, 2025
  • Accepted : November 24, 2025
  • Published : December 31, 2025

Minwook Sua 1 Hyunchul Ahn 1

1국민대학교

Accredited

ABSTRACT

In this paper, we propose a short-term electricity demand forecasting approach that integrates power-related and meteorological data through feature selection and applies a Multi-Head RNN/LSTM/GRU model. By leveraging the strengths of multiple recurrent neural network architectures, the proposed method aims to enhance forecasting accuracy. Furthermore, we fuse power data with meteorological data, identify the combination of features exhibiting the best predictive performance, and apply this to a Multi-Head RNN/LSTM/GRU model. This approach is expected to support informed decision-making in power supply management and contribute to the stable operation of modern power grids.

Citation status

* References for papers published after 2024 are currently being built.