A New Hybrid Model for Hourly Solar Radiation Forecasting Using Daily Classification Technique and Machine Learning Algorithms

Table 3

The seasonal comparison between the ML models.

Error metrics	Winter				Spring				Summer				Autumn
	ANN	SVR	GB	RF	ANN	SVR	GB	RF	ANN	SVR	GB	RF	ANN	SVR	GB	RF
RMSE (W/m^2)	50.06	48.76	50.29	48.94	93.89	94.43	94.27	93.17	41.97	42.01	44.71	43.30	51.42	51.97	51.44	51.22
nRMSE (%)	24.96	24.31	25.07	24.40	22.71	22.84	22.80	22.53	8.40	8.41	8.95	8.67	23.51	23.76	23.52	23.42
MAE (W/m^2)	29.12	27.84	29.45	28.13	59.04	57.09	59.14	56.76	20.24	17.88	22.37	19.34	27.42	27.36	27.35	26.83
nMAE (%)	14.52	13.88	14.68	14.02	14.28	13.81	14.30	13.73	4.05	3.58	4.48	3.87	12.54	12.51	12.51	12.26
AIC	2336	23464	1735	17627	2563	23700	1962	17858	2346	23482	1763	17653	2369	23547	1767	17666