Prediction of Carbon Financial Transaction Price Based on CEEMD Denoising and PSO-LSSVM——Take Guangdong Province as an example

Abstract

Climate change has an unprecedented impact on the world, and it is also an important era proposition to be solved in China's development. Carbon emissions trading occupies an important position, which aims to adapt to the global carbon neutrality context and implement sustainable development requirements. The market mechanism of China is to deal with climate change, and carbon emissions trading occupies an important position. Based on the carbon emission trading price data of Guangdong Province from 2015 to 2021, this paper uses the contribution analysis model of complete ensemble empirical mode decomposition (CEEMD), and particle swarm optimization algorithm-least squares support vector machine (PSO-LSSVM) to predict the carbon price of Guangdong Province. Firstly, the carbon emission trading data of Guangdong Province are obtained. Based on CEEMD decomposition data, the data filtering and noise reduction are completed. The PSO-LSSVM model is used for iterative particles and obtains the optimal solution, and each component is predicted respectively. Finally, the results are integrated. The results are shown that: (1) CEEMD data decomposition noise reduction technology is used to improve the prediction accuracy of IMFs ; (2) The optimization of parameters in LSSVM modeling by PSO algorithm helps to select parameters more reasonably and avoids the randomness of artificial selection to some extent. The research results show that the decomposition ensemble prediction model applies to the carbon emission prediction in Guangdong Province. The good generalization ability of the model provides a more accurate prediction scheme for carbon price prediction in China.