A new circular neural grey model and its application to CO2 emissions in China.

The accurate prediction of China's CO 2 emissions provides a reference and early warning for the implementation of the government's environmental strategy and has far-reaching significance for the world's environmental protection. The carbon emission system is a nonlinear dynamical system influenced by multiple factors with multiple data features. This paper establishes a novel grey differential dynamic multivariate prediction model with recurrent neural network. It uses the grey differential dynamic multivariate prediction model to describe the dynamic characteristics of the nonlinear system. And it uses recurrent neural network to extract data features. The new model is used to compensate for the shortcomings of the grey prediction model that use the least squares method to solve the model parameters affecting the accuracy of the model. The numerical solution of the whitening equations of the new model visualizes the hidden layer state function of the recurrent neural network. The optimal parameters of the new model are obtained through repeated training to improve the prediction accuracy of the model. Finally, the new model is applied to the prediction of China's CO 2 emissions to verify the validity of the new model from four different perspectives. The results show that the simulation and prediction errors of the new model for data of different lengths are less than 5%. The average simulation error in the optimal scheme is 3.01%, and the average prediction error is 0.78%. The results of the new model are better than the results of the other models. This indicates that the new model has better prediction results. Applying the scenario in forecasting the amount of carbon dioxide in China from 2020 to 2026 yields a forecast that matches the actual trend, further illustrating the validity and feasibility of the model. Finally, the forecast results are analysed, and relevant policy recommendations are provided. • A novel neural network grey prediction model is developed. • The model uses neural networks to deeply explore the time-series type data features. • The dynamic nature of the system is reflected using multiple differential grey models. • Successful prediction of CO2 emissions in China. [ABSTRACT FROM AUTHOR]