Author(s)

Keming Cheng, Xinshi Ou, Yangdi Hu

Corresponding Author

Keming Cheng

Abstract

As the global temperature rises, the greenhouse effect becomes more and more obvious, reducing carbon dioxide emissions has become a topic of global concern. Through the background study of carbon emissions, this paper first conducts regional research based on carbon emissions data of China carbon accounting database. Secondly, based on the data of the China Urban Statistical Yearbook, this paper studies the carbon emissions of various regions in the country from 1997 to 2020, explores the spatio-temporal evolution characteristics of each region, and analyzes its influencing factors by constructing the "big data" element momentum entry point. In order to eliminate the error factor as much as possible, a dimension reduction model based on grey correlation analysis was first established to eliminate the factors that have a small impact on carbon emissions. Then, the topsis score comprehensive evaluation model based on entropy weight method is constructed, which shows that carbon emission level is closely related to economy and policy. Finally, SF model, LSTM model, ARIMA time prediction model and SVM model based on factor analysis and genetic algorithm are established to model data indicators, and the optimal model is obtained by comparing the models. Finally, some suggestions on carbon emission are given.

Keywords

carbon emission, Space-time model, TOPSIS legal right Act, SVM model, F model, ARIMA model, LSTM