Analysis of Carbon Emission Influencing Factors and Medium- and Long-term Prediction Methods for Mega Cities Based on GDIM and PSO-XGBoost

In the context of the dual carbon target and to support carbon neutral planning, this paper proposes a Generalized Divisia Index Method(GDIM) and improved EXtreme Gradient Boosting method based on particle swarm optimization(PSO-XGBoost) for analyzing the impact factors and medium- and long-term projections of carbon emissions in mega-cities. First, relevant data, including energy consumption, economic growth rate, electricity production and carbon emission indicators, are collected and the influence contribution is quantified by GDIM method, so as to select the influencing factors with strong correlation. Secondly, methods such as Tapio and decoupling effort model are used to analyze the decoupling trend between economic and other data and carbon emissions, and thus determine the parameter settings under multiple scenarios. Finally, the XGBoost algorithm with particle swarm optimization is used to predict the future carbon emission trends. The actual data of Tianjin, China, verifies the validity of the method and the prediction analysis of the example is carried out.