多标度分形特征下碳排放权价格预测算法

Price forecasting algorithm of carbon emission rights under multiscale fractal characteristics

中文关键词  欧盟碳排放权市场  分形与混沌  小波变换  径向基函数网络  预测分析

英文关键词  EU carbon emissions market  fractal and chaos  wavelet decomposition  radial basis function networks  predictive analytics

中文摘要

      在非线性范式下, 本文构建了基于多贝西小波三层变换和单支重构的遗传算法径向基函数神经网络模型 (daubechies wavelet--genetic algorithm--radial basis function neural network model, Db3--GA--RBF), 探讨了欧盟碳排放 权市场的价格预测问题. 研究表明: 1) 欧盟碳排放权交易市场配额三阶段的现货价格波动均具有局部尺度多样性 特征, 且第3阶段碳价格序列多重分形特征最强, 本质上碳排放权市场是一个多重分形与混沌市场; 2) Db3--GA-- RBF模型能有效地提高数据的准确性和模型的泛化能力, 使模型的预测精度更强; 3) 与其他预测模型效果相比, 基 于施瓦茨信息准则(Schwartz’s information criterion, SIC)的Db3--GA--RBF(SIC)模型的预测精度大约提高70%.

英文摘要

      In this paper, the daubechies wavelet--genetic algorithm--radial basis function neural network (Db3--GA--RBF) model is constructed by nonlinear paradigm, and the price forecasting problem of European Union carbon emissions market (EU--ETS) is discussed. The research showed that: 1) The European Union allowance (EUA) spot price fluctuation in the three stage of the EU carbon emissions market has the characteristics of local scale diversity, and the third stage carbon price series has the strongest multifractal characteristics. Essentially, the carbon emission rights market is a multifractal and chaotic market; 2) The Db3--GA--RBF model can effectively improve the accuracy of data and the generalization ability of the model, and make the prediction accuracy of the model stronger; 3) Compared with other forecasting models, the prediction accuracy of Db3--GA--RBF (SIC) model based on Schwartz’s information criterion (SIC) is improved by about 70%.