Carbon emission forecasting and decoupling based on a combined extreme learning machine model with particle swarm optimization algorithm: the example of Chongqing, China in the “14th Five-Year Plan” period

Abstract

Since the carbon peaking and carbon neutrality goals was included into the ecological civilization construction system, every province and city in China have been actively released their local the carbon peaking and carbon neutrality plans for the “14th Five-Year Plan”. To address the problems of slow updating of carbon emission data and low accuracy of traditional forecasting models, this paper used data from Chongqing, China, to conduct a study on the subject. this paper measured carbon emissions according to the IPCC method，and assessing the development process of resources and environment by means of decoupling analysis. The important factors influencing carbon emissions are selected by the grey correlation method, and the scenario forecast indicators are constructed according to the relevant policy documents of Chongqing, and the important factors and the consumption of coal, oil and natural gas are taken as the inputs of a single forecast model. The following conclusions were obtained: by comparison, the PSO-ELM model is the best model for predicting carbon emissions in Chongqing. The following conclusions were obtained: the combined PSO-ELM prediction model has lower prediction error and higher accuracy, and is more suitable for carbon emission research. The prediction results show that the carbon emissions in Chongqing during the “14th Five-Year Plan” still maintain upward trend, but the growth rate has slowed down compared with 1998-2018, and the carbon emissions tend to stabilize. Overall, there is a weak decoupling between carbon emissions and GDP in Chongqing from 1998 to 2025.

Efficient ELM Model With Parameter Optimization Using Pso Algorithms in the Prediction of Combustion Pressure Parameters of Dsi Engine Using Ethanol- Gasoline Blends

“…Compared to other models, the PSO-ELM yields better prediction performance in terms of the RMSE, MAPE and coefficient of determination (R 2 ) and also PSO-ELM yields prediction values distribution trend as same as that of observed data. Liu B et al, 19 developed PSO-ELM that predict carbon particle emissions in china at Chongqing. They noticed that PSO-ELM model obtains better prediction performance by giving greatest accuracy compared to BP and WOA-BP with minimum MSE, RMSE and MAPE.…”