Forecasting CO 2 emissions in Hebei, China, through moth-flame optimization based on the random forest and extreme learning machine

Abstract

The surge of carbon dioxide emission plays a dominant role in global warming and climate change, posing an enormous threat to the development of human being and a profound impact on the global ecosystem. Thus, it is essential to analyze the carbon dioxide emission change trend through an accurate prediction to inform reasonable energy-saving emission reduction measures and effectively control the carbon dioxide emission from the source. This paper proposed a hybrid model by combining the random forest and extreme learning machine together for the carbon dioxide emission forecasting in this paper; the random forest is applied for influential factors analysis and the extreme learning machine for the prediction. To improve the performance of the prediction model, moth-flame optimization is adopted to optimize initial weight and bias in extreme learning machine. A case study whose data is derived from Hebei Province, China, during the period 1995–2015 is conducted to verify the effectiveness of the proposed model. Results show that the novel model outperforms the compared parallel models in carbon dioxide emission prediction and has the potential to improve the accuracy of CO₂ emission forecasting.