Carbon Price Prediction Based On QRNN Model

The report of the 19 th National Congress of the Communist Party of China regards "persistence of harmonious coexistence between man and nature" as an important part of the basic strategy of adhering to and developing socialism with Chinese characteristics in the new era.Building a "beautiful China" will have a profound impact on China’s future economic growth model.However,the global climate is getting warmer,and greenhouse gas emissions are accumulating,and it is urgent to control greenhouse gas emissions.Therefore,the research on the carbon market has long-term value and important significance,which can enable the participants in the carbon market to better manage risks,give investors scientific decision-making tools,and promote the more stable development of the carbon market.Due to the non-stationary,non-linear,multi-frequency and other irregular characteristics of carbon market prices,it is difficult for traditional models to fully characterize the fluctuations of carbon prices,fully grasp the fluctuation laws of carbon prices,and quantile regression can measure different scores.Predicted value of carbon price under single-digit conditions,which allows investors to better manage risk.Therefore,this paper uses a hybrid model of quantile regression and neural network,which can not only adapt to non-linear sequence analysis,but also fully predict the carbon price fluctuation under different quantile conditions.The results of the study show that:(1)the domestic carbon market is not yet mature,and there are problems such as inconsistent trading volume and inconsistent turnover;(2)the domestic carbon price is affected by domestic and international macroeconomic factors,energy prices,climate,exchange rates and other factors;(3)The digit regression neural network model has better prediction effect than other single models in this study,and the prediction accuracy is higher.The research contribution of this thesis is to expand the theoretical research of domestic carbon price prediction methods,which can provide investors and institutions with scientific decisions and better reference to the changes in the carbon market.