基于CEEMDAN分解和CNN-GRU-Attention模型的碳价格预测研究

Research on Carbon Price Prediction Based on CEEMDAN Decomposition and CNN-GRU-Attention Model

考虑到碳价格数据的波动较大,为更好地帮助企业获取碳排放权,建立起准确的碳交易价格预测模型,本文结合CEEMDAN分解和CNN-GRU-Attention模型来搭建碳价格数据预测模型。在本文提出的方案中,首先对碳价格原始序列进行CEEMDAN分解,通过熵排列重组成新的子序列以提高预测精度、减小计算规模,之后构建CNN-GRU-Attention模型对新的子序列进行预测。通过多模型预测对比可以发现,本文提出的结合CEEMDAN分解和CNN-GRU-Attention模型来搭建的碳价格数据预测模型具有较好的拟合能力,并具备较好的泛化能力。

Considering the significant fluctuations in carbon price data,in order to better assist enterprises in obtaining carbon emission rights and establish an accurate carbon trading price prediction model,this paper combines CEEMDAN decomposition and CNN-GRU Attention models to build a carbon price data prediction model.In the proposed scheme,the original sequence of carbon prices is first decomposed by CEEMDAN,and new subsequences are reorganized through entropy permutation to improve prediction accuracy and reduce computational scale.Then,a CNN GRU Attention model is constructed to predict the new subsequences.Through the comparison of multiple model predic-tions,it can be found that the carbon price data prediction model proposed in this paper,which combines CEEMDAN decomposition and CNN-GRU Attention model,has good fitting ability and generalization ability.