Research On The Influencing Factors And Combination Prediction Model Of Carbon Market Prices In China

At present,the international community has reached a general consensus that excessive carbon dioxide emissions are the main cause of global warming,aggravating the climate system instability,and causing frequent occurrence of extreme weather.Carbon emissions trading system is an important way to deal with climate change and control greenhouse gas emissions.With the proposal of the national goal of "Carbon peak by 2030 and carbon neutral by 2060 in China",domestic and foreign academic fields have shown strong interest on China’s carbon trading system.The carbon emissions allowance trading prices can reflect the demand for emissions reduction,which is a core element of the carbon market operation and a sign of the carbon market maturity,and its fluctuation reflects the scarcity of carbon emissions allowance trading resources.Accurately grasping the characteristics of carbon prices fluctuations can improve the energy consumption of emissions-controlled enterprises and the management capabilities of investors’ carbon assets,and help to promote the realization of emissions reduction goals.In order to clarify the development trend of China’s carbon market and allocate carbon allowances reasonably,this thesis selects 18 influencing factors from the aspects of international carbon assets prices,macroeconomic and industrial development condition,energy prices,exchange rates and climate environment,analyzes the influencing degrees and directions on China’s eight carbon market prices through PSO-ELM-MIV model,and then makes policy recommendations.Next,the non-stationary,nonlinear and chaotic characteristics of eight carbon prices are verified,a VMD-SE-PSO-ELM decomposition integrated combination prediction model based on these characteristics is constructed,which specific steps are as follows: First,utilizing Variational Mode Decomposition(VMD)to process original carbon prices.Secondly,using Sample Entropy(SE)to reconstruct carbon prices components.Then,each carbon prices component is predicted by Extreme Learning Machine(ELM)optimized by Particle Swarm Optimization(PSO).Lastly,the forecasting results of components are added to obtain the final results.Based on the empirical research,the following conclusions are drawn:(1)The same influencing factor has different degrees and directions on China’s eight carbon prices,and the main influencing factors of China’s eight carbon market prices are different,too.Domestic and foreign energy prices,macroeconomic and industrial development condition,and international carbon assets prices have more significant impacts on China’s carbon prices.(2)The proposed VMD-SEPSO-ELM model shows the best prediction performance,fast forecasting speed and has strong accuracy and robustness.(3)The decomposition algorithm effectively improves the prediction accuracy by processing carbon prices with non-stationary,nonlinear and chaotic characteristics into stable and regular subsequences.(4)The PSO-ELM is more suitable for predicting nonlinear and chaotic series,it can improve prediction performance and robustness,and reduce computational complexity.Through the research and relevant conclusions,on the one hand,it can understand the advantages and disadvantages of China’s carbon market structure and mechanism design,and provide policy recommendations for the comprehensive construction and healthy operation.On the other hand,it can also effectively improve the forecasting ability of carbon market participants,so that they can allocate resources reasonably and take appropriate measures to avoid risks.