Carbon Emission Prediction: Methods and Applications Across Fields

The environmental problems hold a close connection to humans daily life, among which the carbon emissions serve an important role. To promote sustainable development, contributions measures need to be conducted to deal with the environmental problems effectively. Currently, abundant studies have discussed the methods to predict carbon emissions so as to help the relevant regulations legislated reasonably and practically. This document focused on the currrent 15 studies regarding carbon emissions prediction methods. The methods examined include traditional statistical models (e.g., grey models, ARIMA), machine learning techniques (e.g., Random Forest, XGBoost), deep learning approaches (e.g., LSTM, CNN), hybrid models (e.g., ARIMA-LSTM, VMD-SSA-LSSVM), and innovative domain-specific methods (e.g., G-code analysis for 3D printing). The results indicate a rather mature studying progress in certain fields, as the individual models, hybrid models and innovative models analyzed all showed great performance and progress to some extent in respective fields compared to specific previous studies.Generally, in spite of the fact that it lacks of a systematic prediction method that is suitable for all diverse industries commonly, various fields are riding on a rising way of improving the methods effectiveness to support the regulation works.