Predictions of carbon emission intensity based on factor analysis and an improved extreme learning machine from the perspective of carbon emission efficiency

Given the severe global warming situation, it is very important to explore the factors influencing carbon emission intensity and accurately analyze the trends in the development of carbon emission intensity to achieve the goal of reducing carbon emissions. In contrast with the existing research, this paper starts from the perspective of carbon emission efficiency, applies stochastic frontier analysis to screen the factors influencing carbon intensity, and constructs a model for predicting carbon emission intensity based on factor analysis and an extreme learning machine. The results suggest that, first, there is a high correlation between carbon emission efficiency and carbon emission intensity. Second, the level of economic development, industrial structure, urbanization level, and government intervention all promote a reduction in carbon emission intensity. The structure of energy consumption and dependence on foreign trade restrain reductions in carbon emission intensity. Finally, the proposed model accurately predicts carbon emission intensity. The research results provide theoretical support for the development of technologies to reduce carbon emissions. This idea can be applied to predict carbon emission intensity in different regions and has practical significance.

Automated summary

This study utilizes stochastic frontier analysis to identify the factors influencing carbon emission intensity and constructs a model for predicting carbon emission intensity based on factor analysis and an extreme learning machine. The results indicate that carbon emission efficiency is highly correlated with carbon emission intensity. Economic development, industrial structure, urbanization level, and government intervention promote a reduction in carbon emission intensity, while energy consumption structure and dependence on foreign trade restrict reductions in carbon emission intensity.