Forecasting Carbon Price with Secondary Decomposition Algorithm and Optimized Extreme Learning Machine

Abstract

Carbon trading is a significant mechanism created to control carbon emissions, and the increasing enthusiasm for participation in the carbon trading market has forced the emergence of higher-precision carbon price prediction models. Facing the complexity of carbon price time series, this paper proposes a carbon price forecasting hybrid model based on secondary decomposition and an improved extreme learning machine (ELM). First, the complementary ensemble empirical mode decomposition with adaptive noise (CEEMDAN) is utilized to decompose the carbon price several intrinsic modal functions to initially weaken the non-linearity of the original carbon price data. Secondly, the first intrinsic mode function (IMF1) with the strongest volatility is processed by the variational mode decomposition (VMD). Then, the partial autocorrelation function (PACF) is applied to obtain the model input variables for subsequences. Finally, the ELM improved by the bald eagle search (BES) algorithm is utilized to make predictions. In the empirical analysis, five actual datasets from three carbon markets are used to verify the prediction performance of the proposed model. Based on the six evaluation indicators of the predicted results, the proposed model is the best performer among all models, which suggests that CEEMDAN-VMD-BES-ELM is effective and stable in predicting carbon price.

Global Trends of Carbon Finance: A Bibliometric Analysis

“…In this cluster, researchers focus on improving the accuracy and stability of carbon price forecasting. Moreover, they have proposed a number of carbon price forecasting model approaches, such as the deep neural network model TCN-Seq2Seq [97], a hybrid model for carbon price forecasting based on quadratic decomposition and improved extreme learning machine (ELM) [98], a hybrid model of GARCH and long-term short-term memory network [99], a quadratic decomposition carbon price forecasting model based on a kernel limit learning machine optimized by the sparrow search algorithm [100], and other related carbon price forecasting models. The main keywords included in the first six clusters are listed in Table 6.…”

A Novel Hybrid Model for Forecasting China Carbon Price Using CEEMDAN and Extreme Learning Machine Optimized by Whale Algorithm

“…Conducted the EMD technology to decompose the carbon price, used the particle swarm optimization least squares support vector machine (PSO-LSSVM) model for out-of-sample forecasting, the results suggest the EMD-PSO-LSSVM model has forecasting superiority in Europe carbon price [4][5].To reduce the decomposition noise, the CEEMDAN and VMD technologies were used to perform primary and secondary decomposition of the original price, LSTM and ELM models were used for forecasting. The findings put that the CEEMDAN-VMD-LSTM and CEEMDAN-VMD-ELM models have substantially forecasting accuracy in China carbon market [6][7][8].…”

Forecasting China carbon price using an error-corrected secondary decomposition hybrid model integrated fuzzy dispersion entropy and deep learning paradigm

“…For example, employed the EMD, EEMD, CEEMD, CEEMDAN and VMD technologies to primary decompose and secondary decompose the original carbon price, further built the deep learning models such as PSO-LSSVM, BP network, gate recurrent unit (GRU) , LSTM and ELM improved by the vulture search (BES) algorithm for the out-of -sample forecasting (Li et al2021;Wang et al2022), the results convinced that the EMD-VMD-BP and EMD-VMD-LSTM model have relatively stable performance performance in Beijing and Shanghai carbon markets (Sun and Huang 2020), while VMD-EEMD-GRU, CEEMD-VMD-BPNN and CEEMDAN-VMD -LSTM model have superior multi-step forecasting accuracy in Hubei and Guangdong carbon markets (Wu and Liu 2020;Zhou and Wang 2021;Liu et al2023). Theoretically, ICEEMDAN technology has strong signal decomposition advantages (Liu et al2023).…”