Temperature Forecasting via Convolutional Recurrent Neural Networks Based on Time-Series Data

Table 2

Performance comparison with different approaches.

Approaches	MAE	RMSE
Operational consensus forecasts	1.253	2.174
Original backpropagation neural networks	0.934	1.952
Stacked denoising autoencoder	0.951	1.863
Convolutional recurrent neural networks	0.907	1.697