Forecasting carbon dioxide emissions using adjacent accumulation multivariable grey model

Carbon dioxide (CO2) emissions, the primary greenhouse effect catalyst, command global attention due to associated environmental challenges. Urgent carbon reduction is imperative, particularly with scholarly discourse emphasizing the criticality of peak emissions and carbon neutrality. Accurate CO2 emission prediction holds immense importance for shaping effective management policies aimed at emission reduction and environmental mitigation. This study introduces an enhanced multivariable grey prediction model (AGMC(1,N)), utilizing the particle swarm optimization (PSO) algorithm based on artificial intelligence to determine its optimal order. Rigorous analysis, including a disturbance bound classification discussion, validates the superior stability and outstanding predictive capability of the AGMC(1,N) model, as exemplified in a detailed case study. Applying the AGMC(1,N) model to forecast CO2 emissions in the Beijing-Tianjin-Hebei region and Shanxi Province reveals a correlation between energy, primary and secondary industry growth, GDP per capita, and increased emissions, while rising urbanization and natural gas consumption correlate with emission decline. The study concludes with actionable proposals derived from predictive insights, providing valuable support for decision-making by management departments focused on emission reduction. (c) 2024 International Association for Gondwana Research. Published by Elsevier B.V. All rights are reserved, including those for text and data mining, AI training, and similar technologies.