Carbon emissions prediction based on the GIOWA combination forecasting model: A case study of China

Mitigating greenhouse gas emissions is a significant global challenge, and precise prediction of carbon emissions holds the utmost importance for China to attain its dual carbon goal. To overcome the limitations of existing studies using a single forecasting method, based on quadratic exponential smoothing, multiple linear regression, and Gaussian process regression models, this paper constructs a carbon emission combination forecasting model with the generalized induced ordered weighted average (GIOWA) operator and analyzes carbon emission reduction performance. Empirical testing utilizing China's carbon emission data from 1980 to 2020 reveals the following findings: (1) The GIOWA combination forecasting model significantly enhances the accuracy of carbon emission forecasts, with an average accuracy exceeding 99.5% over the sample period, surpassing various single forecasting methods. (2) The carbon emission reduction target can be achieved under three different scenarios of low, average, and high GDP growth rates. Specifically, under the low growth rate scenario, China can achieve a 60% reduction in carbon intensity by 2025 and a 65% reduction by 2030. This study offers valuable decision support for the development of effective carbon reduction policies.