A Hybrid Carbon Price Forecasting Model with External and Internal Influencing Factors Considered Comprehensively: A Case Study from China

Abstract

With the continuous emission of greenhouse gases, the carbon trading market has become a powerful weapon to contain it. It is indispensable to analyze the carbon price of China that acts as the largest emitter of carbon dioxide worldwide. Therefore, this paper proposes an innovative hybrid carbon price forecasting model that incorporates fast ensemble empirical mode decomposition (FEEMD) and extreme learning machine optimized by particle swarm optimization (PSO-ELM) with external and internal influencing factors considered. The original carbon price series are disassembled into several intrinsic mode functions (IMFs) and one residual via FEEMD. The PSO-ELM is then employed to forecast the sub-series. It's remarkable that the inputs of the PSO-ELM model are divided into external and internal influencing factors. Factor analysis is used to extract potential factors from energy prices, macroeconomics and other influencing factors associated with the original carbon price as external influencing factors, and the partial autocorrelation function (PACF) is exploited to select internal influencing factors. A case study in Hubei Province, China shows that the proposed carbon price forecasting model is superior to the contrast models in terms of the smallest prediction error (MAE = 0.1274 yuan, MAPE = 0.8368%) and the strongest stability (RMSE = 0.0116 yuan). And the forecasting results demonstrate that the developed model with external and internal influencing factors considered can highly improve carbon price prediction performance and have potential in a wider range of carbon price forecasting. In addition, accurate carbon price forecasting can help the government realize macro control and the investors fulfill risk minimization.

A Study on the Implementation of Corporate Carbon Emission Accounting in the Production Process

“…The historical cost method is difficult to reflect the price fluctuations of carbon emission rights in the carbon trading market promptly [38]. With the implementation of the "Measures for the Administration of Carbon Emission Trading (for Trial Implementation)", China's carbon trading market is becoming increasingly mature, and the market quotations for carbon emission rights are relatively easy to obtain.…”

Forecasting Premium Motor Spirit (PMS) and Energy Commodities Prices Using Machine Learning Techniques: A Review

“…8. (Sun et al, 2020) Carbon price forecasting Fast ensemble empirical mode decomposition (FEEMD) and extreme learning machine optimized by particle swarm optimization (PSO-ELM)…”

“…As a result, there is a need to evolve more accurate forecasting models. Similarly, extreme learning machine model incorporated with Fast Ensemble Empirical Mode Decomposition (FEEMD) optimized by Particle Swarm Optimization (PSO-ELM) (Sun, et al, 2020) and extreme learning machine, improved grey model and Hodrick-Prescott (Zhao, et al, 2021) was proposed as a hybrid decomposition and integration prediction model for carbon price. But future work is required to explore the energy pricing regime in determining the effectiveness of the proposed model.…”

The Influencing Factors of the Carbon Trading Price: A Case of China against a “Double Carbon” Background

“…But these factors differ in factor selection, time, or spatial dimension and are not necessarily leading to consistent conclusions [9]. Much more literature focuses on the European carbon market or individual pilot market in China [4,[10][11][12][13][14][15][16], while there are few literature studies for the 2017-2021 period of the China markets. Factors vary according to single or mixed types, and the corresponding impacts empirically differ in their correlation direction and degree across pilot markets [1,4,15].…”