Using the artificial neural networks for prediction and validating solar radiation

Journal of the Egyptian Mathematical Society

Table 2 The statistical data for training and testing of the first method

Cites	Model	η	RMSE	MِAPE	MABE	r	R²	Rank
For training
Cairo	1	0.01	0.977	4.256	0.757	0.9868	0.9996	1
	2	0.2	1.184	5.431	0.943	0.9809	0.9996	2
	3	0.3	0.977	4.268	0.760	0.9869	0.9999	1
Borg El-Arab	1	0.01	1.307	6.315	1.028	0.9826	0.9998	3
	2	0.2	1.025	4.746	0.772	0.9893	0.9999	2
	3	0.3	1.019	4.699	0.765	0.9894	0.9999	1
Aswan	1	0.01	1.444	5.270	1.108	0.9659	0.9998	3
	2	0.2	1.425	5.332	1.105	0.9685	0.9999	2
	3	0.3	1.324	5.223	1.103	0.9687	0.9999	1
For testing
Cairo	1	0.01	3.117	14.106	2.396	0.8584	0.9998	3
	2	0.2	2.872	12.680	2.141	0.8898	0.9998	2
	3	0.3	2.662	11.852	2.026	0.8990	0.9999	1
Borg El-Arab	1	0.01	3.484	15.925	2.612	0.8817	0.9997	3
	2	0.2	2.986	12.875	2.095	0.9261	0.9997	2
	3	0.3	2.824	12.175	2.144	0.9283	0.9999	1
Aswan	1	0.01	1.713	6.362	1.294	0.9403	0.9998	3
	2	0.2	1.656	5.948	1.175	0.9509	0.9998	2
	3	0.3	1.555	5.742	1.073	0.9609	0.9999	1