A novel deep learning based forecasting model for carbon emissions trading: A comparative analysis of regional markets

Abstract

This paper proposes a deep learning-based forecasting system for carbon emissions trading in regional markets. The system utilizes a hybrid convolutional neural network (CNN) deep learning algorithm to predict future carbon emission levels and facilitate the trading of carbon emission allowances. The system is compared to traditional methods of forecasting in order to assess the accuracy and performance of the system. In order to enhance the CNN performance, a new optimization algorithm based on bacteria foraging algorithm (BFA) is proposed which uses a modification to make a global search. By leveraging digital twins in the markets, a comparison is conducted using data from three regional carbon markets: the European Union Emissions Trading System, the Regional Greenhouse Gas Initiative, and the China Carbon Market. Results show that the proposed BFA-CNN based deep learning-based system outperforms traditional forecasting methods in terms of accuracy and provides more reliable estimates of future carbon emissions. The proposed system is a novel approach to carbon emissions trading and has the potential to improve the efficiency of regional carbon markets.