基于GRA-LSTM模型的碳排放量预测——以宿迁市为例

Carbon Emission Forecast Based on GRA-LSTM Model——Through Taking Suqian City as an Example

为提高碳排放量预测精度,构建了一类基于GRA-LSTM神经网络的碳排放量预测模型。借助该模型,对江苏省宿迁市的碳排放量进行预测仿真实验,并将预测结果与BP神经网络、一阶一元灰色GM(1,1)模型的预测结果进行比较。结果表明,GRA-LSTM模型的预测误差RMSE、MAE、MAPE分别为0.0665、0.1996、0.0665,具有较高的预测精度。

In order to improve the accuracy of carbon emission forecast,the study constructs carbon emission forecast model based on GRA-LSTM neural network.With the help of the model,the study carries out the simulation experiment of carbon emission in Suqian City,Jiangsu Province,and compares the forecast results with the forecast results of BP neural network,first-order unitary gray GM(1,1)model.The results show that RMSE,MAE and MAPE of forecast errors of GRA-LSTM model are 0.0665,0.1996 and 0.0665,with higher forecast accuracy.