Influential Factor Analysis and Projection of Industrial CO_2 Emissions in China Based on Extreme Learning Machine Improved by Genetic Algorithm

摘要

In China, CO2 emissions from industrial sectors on a larger scale than other end-use sectors. In order to reduce CO2 emissions, it is necessary to study the influencing factors and projection of industrial CO2 emissions. Based on accounting for CO2 emissions from the industrial sectors, this paper carries out bivariate correlation analysis and linear regression analysis on 15 preselected influencing factors and industrial CO2 emissions, removing two factors that have failed the significance test. In order to obtain some potential commonalities among the influencing factors, the remaining 13 influencing factors are divided into four categories, and then factor analysis is performed on each category in order to obtain five latent factors. An extreme learning machine algorithm that uses genetic algorithms to optimize the input weights and bias thresholds - the genetic algorithm extreme learning machine (GA-ELM) algorithm - to predict industrial CO2 emissions, the empirical results show that the GA-ELM algorithm using five factors as inputs has a higher prediction accuracy and performance for industrial CO2 emissions than the extreme learning machine, back propagation neural network, and back propagation neural network optimized by the genetic algorithm. It also shows that the five influencing factors have a significant impact on industrial CO2 emissions. Finally, based on the analysis of five influencing factors, some policy recommendations are proposed for the CO2 emissions reduction path in the industrial sectors.