China's carbon dioxide emission forecast based on improved marine predator algorithm and multi-kernel support vector...

Area of Science:

• Environmental Science
• Climate Change Research
• Computational Intelligence

Background:

• Global warming is a significant environmental issue driven primarily by carbon dioxide emissions from fossil fuels.
• Accurate forecasting of carbon dioxide emissions is crucial for developing effective climate change mitigation strategies.
• Existing forecasting models may lack the accuracy required for precise policy-making.

Purpose of the Study:

• To propose a novel hybrid forecasting model for carbon dioxide emissions.
• To enhance prediction accuracy through an improved optimization algorithm.
• To forecast China's carbon dioxide emissions during the "14th Five-Year Plan" period.

Main Methods:

• Development of an enhanced marine predator algorithm (EGMPA) incorporating elite opposition-based learning and the golden sine algorithm.
• Integration of the EGMPA with a multi-kernel support vector regression model.
• Validation using China's carbon dioxide emission data from 1965 to 2020, evaluated by RMSE, MAE, and MAPE.

Main Results:

• The EGMPA demonstrated improved convergence speed and optimization accuracy compared to the standard MPA.
• The proposed hybrid model achieved high prediction accuracy with RMSE of 37.43 Mt, MAE of 30.63 Mt, and MAPE of 0.32%.
• The model accurately predicted a continued increasing trend in China's carbon dioxide emissions during the "14th Five-Year Plan" period, albeit with a slowing growth rate.

Conclusions:

• The proposed EGMPA-multi-kernel support vector regression model offers a significant improvement in carbon dioxide emission forecasting accuracy.
• The model provides a reliable tool for predicting future carbon dioxide emissions, aiding in climate change policy development.
• China's carbon dioxide emissions are projected to increase but at a decelerating rate during the "14th Five-Year Plan" period.