Explainable Dual-Attention Encoder–Decoder Model for Natural Gas Consumption Forecasting Using Algerian Hourly Data

Randa Ladlani (), Samiha Ait Taleb () and Abderrazak Sebaa ()
Additional contact information
Randa Ladlani: École supérieure en Sciences et Technologies de l’Informatique et du Numérique, Laboratoire LITAN
Samiha Ait Taleb: École supérieure en Sciences et Technologies de l’Informatique et du Numérique, Laboratoire LITAN
Abderrazak Sebaa: École supérieure en Sciences et Technologies de l’Informatique et du Numérique, Laboratoire LITAN

A chapter in Proceedings of the International Conference on Artificial Intelligence Applications in Business Administration in MENA Region (ICAIABA 2026), 2026, pp 345-355 from Springer

Abstract: Abstract Natural gas consumption forecasting supports efficient energy management and resource planning in industrial environments. The proposed architecture combines an encoder–decoder structure with dual attention mechanisms — temporal and feature-level — alongside cyclical encodings and moving-average representations to improve forecasting accuracy and stability. Evaluation on Algerian hourly natural gas data yields Test MAE = 0.0255 and R2 = 0.9740, outperforming classical machine learning and deep learning baselines by up to 38%. Cross-domain validation on the GEFCom2014 electricity load benchmark confirms generalizability (R2 = 0.9817, 67% MAE reduction over XGBoost). SHAP analysis quantifies feature contributions, identifying historical consumption and temporal encodings as the dominant predictors, with meteorological variables providing secondary refinement.

Keywords: Natural Gas Consumption Forecasting; Dual Attention Mechanism; Encoder-Decoder Architecture; Explainable AI; SHAP (search for similar items in EconPapers)
Date: 2026
References: Add references at CitEc
Citations:

There are no downloads for this item, see the EconPapers FAQ for hints about obtaining it.

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:spr:advbcp:978-94-6239-711-8_32

Ordering information: This item can be ordered from
http://www.springer.com/9789462397118

DOI: 10.2991/978-94-6239-711-8_32

Access Statistics for this chapter

More chapters in Advances in Economics, Business and Management Research from Springer
Bibliographic data for series maintained by Sonal Shukla () and Springer Nature Abstracting and Indexing ().