Combined Long Short-Term Memory Network-Based Short-Term Prediction of Solar Irradiance

Table 3

Comparative analysis of the proposed CLSTMN with the baseline prediction model on sunny and cloudy datasets.

S. No	Authors	Year	Prediction model	Evaluation metric (RMSE in W/m^2)
				Sunny days	Cloudy days
1	Xiaoyan Xiang et al.	2021	Persistence	48.9464	49.1759
2	Ferrari, Stefano et al.	2013	ARIMA	3.0143	3.1354
3	de Araujo, Jose Manuel Soares	2020	WRF	25.4847	26.2691
4	Mishra, Sakshi & Praveen Palanisamy	2018	RNN	1.4719	2.0143
5	Kartini, Unit Three & Chao Rong Chen	2017	k-NN-BPLNN	0.9464	0.9870
6	Kuk Yeol Bae et al.	2017	SVM	0.7997	0.8050
7	Omaima El Alani et al.	2019	MLP	0.4041	0.4357
8	Yazeed A. A-Sbou & Khaled M. Alawasa	2017	NARX	0.5268	0.5794
9	Naylani Halpern-Wight et al.	2020	LSTM	0.2519	0.3133
10	Manoharan Madhiarasan & Mohamed Louzazni	2022	Proposed CLSTMN 1 hour ahead
			Proposed CLSTMN 2 hours ahead	0.0031	0.0069
			Proposed CLSTMN 3 hours ahead	0.0087	0.0273
			Proposed CLSTMN 4 hours ahead	0.0130	0.0328
			Proposed CLSTMN 5 hours ahead	0.0162	0.0029
			Proposed CLSTMN 6 hours ahead	0.0157	0.0176

Bold implies the best result.