A STUDY ON CARBON PEAKING PATHS IN HENAN, CHINA BASED ON LASSO REGRESSION-BP NEURAL NETWORK MODEL

To explore the path to carbon emission peaking of Henan, China and meet the strategic need of the local authorities, in this paper, the data of twelve factors of social, economic, energy consumption, and resources in Henan from 2001 to 2020 were selected, and the Lasso-BP neural network method was used to establish a prediction model of carbon emission in Henan province. Based on the regression analysis of the data of the 12 indexes, six different development paths were designed to predict the carbon emission of Henan from 2021 to 2035. The results showed that: 1) among the twelve factors, six key factors affecting carbon peaking were the share of coal consumption, energy consumption per unit of GDP, forest coverage, total energy consumption, the share of secondary industry GDP, and private car ownership; 2) paths 1 to 4 pursuing single-factor development were all unable to achieve carbon peaking in 2030. Under paths 5 and 6, Henan would reach peak carbon in 2029. Compared with path 5, the peak CO₂ emission of path 6 was 2.53 Mt lower, with a peak of 510.91 Mt CO₂; 3) to achieve the carbon emission peak, the average annual growth rates of coal consumption, energy consumption per unit of GDP, forest coverage, total energy consumption, the share of secondary industry in GDP, and private car ownership should be controlled at -4.0% and -5.0%, -3.5% and -4.0%, 2.0% and 3.0%, 0.5% and 0.4%, -1.5% and -2.0%,7.5% and 7.0%, during the 14th Five-Year Plan and the 15th Five-Year Plan Period respectively.