Topological Attention and Deep Learning Integration for Electricity Consumption Forecasting

Ahmed Ben Salem, Manar Amayri

2025

Abstract

In this paper, we consider the problem of point-forecasting of univariate time series with a focus on electricity consumption forecasting. Most approaches, ranging from traditional statistical methods to recent learning-based techniques with neural networks, directly operate on raw time series observations. The main focus of this paper is to enhance forecasting accuracy by employing advanced deep learning models and integrating topological attention mechanisms. Specifically, N-Beats and N-BeatsX models are utilized, incorporating various time and additional features to capture complex nonlinear relationships and highlight significant aspects of the data. The incorporation of topological attention mechanisms enables the models to uncover intricate and persistent relationships within the data, such as complex feature interactions and data structure patterns, which are often missed by conventional deep learning methods. This approach highlights the potential of combining deep learning techniques with topological analysis for more accurate and insightful time series forecasting in the energy sector.

Paper Citation

in Harvard Style

Ben Salem A. and Amayri M. (2025). Topological Attention and Deep Learning Integration for Electricity Consumption Forecasting. In Proceedings of the 14th International Conference on Smart Cities and Green ICT Systems - Volume 1: SMARTGREENS; ISBN 978-989-758-751-1, SciTePress, pages 15-23. DOI: 10.5220/0013116800003953

in Bibtex Style

@conference{smartgreens25,
author={Ahmed Ben Salem and Manar Amayri},
title={Topological Attention and Deep Learning Integration for Electricity Consumption Forecasting},
booktitle={Proceedings of the 14th International Conference on Smart Cities and Green ICT Systems - Volume 1: SMARTGREENS},
year={2025},
pages={15-23},
publisher={SciTePress},
organization={INSTICC},
doi={10.5220/0013116800003953},
isbn={978-989-758-751-1},
}

in EndNote Style

TY - CONF

JO - Proceedings of the 14th International Conference on Smart Cities and Green ICT Systems - Volume 1: SMARTGREENS
TI - Topological Attention and Deep Learning Integration for Electricity Consumption Forecasting
SN - 978-989-758-751-1
AU - Ben Salem A.
AU - Amayri M.
PY - 2025
SP - 15
EP - 23
DO - 10.5220/0013116800003953
PB - SciTePress