Variable Split Convolutional Attention: A novel Deep Learning model applied to the household electric power consumption

Rui Gonçalves, Vitor Miguel Ribeiro and Fernando Lobo Pereira

Energy, 2023, vol. 274, issue C

Abstract: The accurate prediction of electric power consumption in the residential sector is a desirable action to minimize potential energy losses and maximize social welfare. The goal of this study is to propose a new Deep Learning Neural Network architecture for multivariate time series problems, which includes a novel attention mechanism applied to the Convolutional Long Short-Term Memory Network model. The new attention mechanism is implemented with convolutional layers, splits the data by explanatory variable, incorporates the cyclical segmentation of data by day, and uses causal and roll padding to ensure proper information augmentation before convolutional operations. The output of the attention block is a bi-dimensional context map for each explanatory variable. Considering the Household Electric Power Consumption data set provided by the repository of the University of California at Irvine, the proposed Variable Split Convolutional Attention model is trained, tested, and compared with several alternatives. The main result of this study reveals that the innovative model exhibits the lowest forecasting error.

Keywords: Electric energy consumption; Deep learning; Attention mechanism (search for similar items in EconPapers)
Date: 2023
References: View references in EconPapers View complete reference list from CitEc
Citations: View citations in EconPapers (3)

Downloads: (external link)
http://www.sciencedirect.com/science/article/pii/S0360544223007156
Full text for ScienceDirect subscribers only

Related works:
This item may be available elsewhere in EconPapers: Search for items with the same title.

Export reference: BibTeX RIS (EndNote, ProCite, RefMan) HTML/Text

Persistent link: https://EconPapers.repec.org/RePEc:eee:energy:v:274:y:2023:i:c:s0360544223007156

DOI: 10.1016/j.energy.2023.127321

Access Statistics for this article

Energy is currently edited by Henrik Lund and Mark J. Kaiser

More articles in Energy from Elsevier
Bibliographic data for series maintained by Catherine Liu ().