Complementary ARIMA and XGBoost carbon emission timing forecasting methods for different terminal energy consumption industries

“Carbon peak” and “carbon neutrality” are the major strategic decisions put forward by China as the largest developing country to cope with global temperature rise. Studying and predicting carbon dioxide emissions of different terminal energy consumption industries was an important work in carbon emission reduction. Based on the data of different terminal energy consumption industries in a central province, ARIMA model and XGBoost model were respectively used to forecast the carbon emissions of different terminal energy consumption industries in time series, and the outliers were analyzed. The results showed that the accuracy of time series prediction can be optimized by adjusting the model parameters, while ARIMA model was not suitable for the time series prediction of carbon emissions of the energy production and processing conversion industry. The XGBoost model was not suitable for the carbon emission time series prediction of residential life. The main reason was that the carbon emission time series prediction of energy production and processing conversion industry was nonlinear, while the ARIMA model mainly studies linear problems. And the carbon emission time series prediction of residential life has prominent linear characteristics, which cannot be described by non-linear model, resulting in a large deviation in the prediction results. According to the results of this paper, the carbon emissions of seven terminal energy consumption Sectors can be time series predicted by ARIMA model and XGBoost model in complementary, and the MPAE values between the real values and the predicted values are all much lower than 20%. Among them, agriculture, industry, construction, transportation and service industries and other industries can be predicted more accurately by using the two methods. Above all, it provides more accurate guidance for the carbon emission reduction policy of the whole field of terminal energy consumption industry.