Attention-Based Deep Learning for CO2 Global Emission Forecasting to Support Sustainable Global Climate Actions

Matta, Akhil, Stynes, Paul and Muntean, Cristina Hava (2026) Attention-Based Deep Learning for CO2 Global Emission Forecasting to Support Sustainable Global Climate Actions. In: 2026 International Conference on Artificial Intelligence for Sustainable Engineering and Innovation, AISEI 2026. IEEE, Irbid, Jordan, pp. 927-932. ISBN 979-833157976-0

Full text not available from this repository.

Abstract

Accurately forecasting CO2 emissions remains a significant research challenge due to the complex and heterogeneous patterns that emerge across different countries, economic sectors, and time periods. This study addresses this challenge by analyzing a large, multi-sector global emissions dataset and evaluating a diverse set of machine learning and deep learning models within a unified predictive framework. The evaluated approaches include ensemble-based models such as Random Forest, Gradient Boosting, and XGBoost, deep learning architectures including Deep Neural Networks and Long Short-Term Memory (LSTM) networks and attention-based tabular model (TabNet). The methodology incorporates comprehensive data preparation, temporal feature engineering, and systematic comparative evaluation across all models. The findings demonstrate that attention-based tabular learning provides substantial gains in predictive accuracy for global emission forecasting. Among all evaluated models, TabNet emerged as the best-performing approach, achieving an RMSE of 0.490 and an R2 of 0.993, significantly outperforming traditional ensemble methods and sequence-based deep learning models in capturing multi-region and multi-sector CO2 emission dynamics. The outcomes of this research further support global sustainability efforts by contributing to UN SDG 7 (Affordable and Clean Energy), SDG 9 (Industry, Innovation and Infrastructure), SDG 11 (Sustainable Cities and Communities), and SDG 13 (Climate Action) through enhanced emission monitoring and decision-support capabilities.

Item Type:	Book Section
Uncontrolled Keywords:	CO2 Emissions Forecasting; Deep Learning; Global Emissions Modelling; Sustainable climate actions
Subjects:	G Geography. Anthropology. Recreation > GE Environmental Sciences G Geography. Anthropology. Recreation > GE Environmental Sciences > Environmental protection > Climate change mitigation Q Science > Q Science (General) > Self-organizing systems. Conscious automata > Machine learning
Divisions:	School of Computing > Staff Research and Publications
Depositing User:	Tamara Malone
Date Deposited:	24 Jul 2026 11:54
Last Modified:	24 Jul 2026 14:29
URI:	https://norma.ncirl.ie/id/eprint/9480

Actions (login required)

View Item