Carbon price forecasting with optimization prediction method based on unstructured combination

The construction of carbon emission trading market is gradually improved, making carbon assets have financial nature, which can effectively restrain excessive carbon emissions. Accurate prediction of the carbon price is of great significance to the scientific decision-making of the government. In order to make the prediction more accurate and reasonable, this paper proposes a new combinatorial optimization prediction method based on unstructured data. In the model, firstly, the structured data screened by grey correlation method and factor analysis and the unstructured data screened by Baidu index are taken as one of the input ends of prediction. Secondly, theMean value Optimization (MOEMD) method is used to decompose the fluctuating carbon price as the other part of the input of the prediction model. Then, based on the optimized Extreme LearningMachine (ELM) prediction model, the Kidney Algorithm (KA) algorithm with scaling factor and cooperation factor (CKA) model are established to predict the carbon trading price of China. Finally, simulation experiments are carried out in eight pilot areas in China to verify the effectiveness of the model. The results show that the MOEMD-CKA-ELM performs well in carbon price prediction, and the unstructured learning method effectively improves the prediction performance of the model. (C) 2020 Elsevier B.V. All rights reserved.