| Regional carbon reduction related work smoothly and sustainable development,cannot leave the scientific and reasonable prediction of carbon dioxide emissions.With the development of application,the related carbon emission prediction research also needs to be continuously promoted systematically.Therefore,this paper systematically analyzes the carbon dioxide emissions in the regional environment from three dimensions of country,province and city.Based on the national energy consumption data,this paper analyzes the national carbon dioxide emissions and observes their changing trend.Compared with the traditional ARIMA model,GM(1,1)model,SVR and GBDT,the results show that the prediction effect of machine learning model is better than the two traditional models.Further weighted combination of SVR and GBDT according to the prediction results.The prediction results show that the combination model has the highest prediction accuracy.According to the forecast results of this model,it is found that the national carbon dioxide emissions will have a slight downward trend in the next few years.Then,in order to better understand the situation of carbon dioxide emissions in China,this paper conducts K-Means clustering for provinces based on provincial data,it is found that the clustering effect is better by referring to the four features of BE,CID,NM and MAC,which has important reference significance for the time series clustering problem.According to the clustering results,one province from each category was selected for prediction,and the comparison results of the models further demonstrated the applicability of the combination model in the prediction of carbon dioxide emissions.Then,this paper further analyzes the carbon emissions of specific provinces from the city level.By ranking the influencing factors of carbon emissions,it is found that the construction industry and tertiary industry are the main factors affecting carbon emissions in this province,which can serve as a warning for the local government when adjusting the emission reduction target. |
