Prediction of national carbon emission efficiency based on ARMI-Convolution LSTM*

This is the authors' abstract. We don't add key points for this paper.

In this paper, convolutional long and short term neural network model(LSTM) and ARIMA model are used to fit and forecast China's carbon emission efficiency, and the weights of the two in the combined prediction are set to different values to observe the optimal fitting results and predict the future trend. First of all, in terms of prediction accuracy, the results show that the prediction accuracy of the long-term and short-term neural network model and the time series ARIMA model is less than the combined prediction accuracy. Compared with each single prediction, the combined prediction not only considers the complementarity of linear and nonlinear, but also considers the combination of mathematical model prediction and influencing factor prediction, which makes the prediction result more practical significance. Finally, the analysis results have certain enlightenment significance for the future trend.