Short-term forecasting of surface solar incident radiation on edge intelligence based on AttUNet

Journal of Cloud Computing

Advances, Systems and Applications

Table 3 Comparison of all performance indicators for spring, summer, and autumn 2021-2022

Season	Model	MAE	RMSE	nRMSE	SSIM
Spring	ConvLSTM	118.25	171.49	39.80	0.71
	PhyDNet	107.01	160.38	36.21	0.76
	Traj-GRU	111.24	165.91	38.04	0.74
	E3D-LSTM	108.14	160.33	36.04	0.77
	PredRNN	101.78	155.05	34.26	0.78
	PredRNN++	100.78	153.05	32.26	0.80
	EDA-AttUNet	97.22	147.71	32.55	0.81
Summer	ConvLSTM	112.11	167.28	38.12	0.74
	PhyDNet	104.37	151.29	34.83	0.83
	Traj-GRU	106.33	157.44	35.14	0.80
	E3D-LSTM	104.24	155.91	37.32	0.76
	PredRNN	93.61	135.71	29.55	0.87
	PredRNN++	91.46	133.22	28.43	0.89
	EDA-AttUNet	94.20	149.05	31.16	0.84
Autumn	ConvLSTM	109.48	166.44	37.32	0.74
	PhyDNet	99.88	148.65	32.35	0.82
	Traj-GRU	103.71	153.29	34.04	0.81
	E3D-LSTM	101.24	147.91	33.01	0.79
	PredRNN	91.98	144.01	32.15	0.83
	PredRNN++	89.28	140.21	31.33	0.85
	EDA-AttUNet	87.27	130.22	27.15	0.88