Modeling Energy-Related CO₂ Emissions with Backpropagation and Metaheuristics

In this paper, we propose a model that combines Backpropagation Neural Network (BPNN) with a metaheuristic algorithm to predict carbon dioxide (CO <inf xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">2</inf> ) emissions. The model utilizes input variables directly influencing carbon dioxide emissions. Our objective is to monitor CO <inf xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">2</inf> emissions based on energy consumption resulting from production processes driven by the corresponding energy demand. Model training and testing are conducted to identify the most suitable model. The results demonstrate that machine learning with the Particle Swarm Optimization (PSO) model achieves an optimal avoidance rate. To assess model accuracy using Mean Square Error (MSE) and Mean Absolute Error (MAE) the smallest error indicates a more precise prediction. Following established guidelines, this document aids managers in making informed business decisions and devising energy management policies that effectively monitor carbon emissions.