Table 2.

A summary of the studies regarding the prediction of solar radiation data using time series methods.

Location	Model	Input Parameters	Output Parameter	Data Scale	Statistical Benchmarks	Key Findings	References
Kansas, Denver, and Arizona, United States	ARIMA	Historical solar radiation	Daily average global horizontal solar radiation	January 01, 1987, to December 31, 1990	MAPE	ARIMA has better results when applied to the time-varying daily solar radiation prediction.	[52]
Many countries in Europe and Asia	ARIMA	Historical solar radiation	Daily solar radiation	9436 days from November 16, 1978	N/A	The time series in different climatic conditions depend on the long-range variability of solar radiation.	[53]
Jeddah, Saudi Arabia	AR	Historical solar radiation	Diffuse horizontal radiation, hourly global solar radiation, and direct normal radiation	1998–2002	RMSE and MBE	The applicable model developed with the entered parameters resulted in an accurate prediction.	[54]
Miami and Orlando, United States	ARIMA	Historical solar radiation	Hourly global solar radiation	January 1995 to December 2005	MBE and RMSE	It was observed that cloud cover information gives more accurate results in terms of prediction.	[55]
Awali, Bahrain	ARIMA	Historical solar radiation	Daily average solar radiation	May 2010 to April 2011	MAPE	It was observed that the different ARIMA models used effectively predict solar radiation.	[56]
Five cities in France	ARMA	Meteorological parameters	Hourly global solar radiation	October 2002 to December 2008	nRMSE	The nRMSE ranges from 18.9 to 21.1% in five different cities.	[57]
Ajaccio, France	ARMA	Meteorological parameters	Hourly solar radiation	2007–2008	nRMSE	ANN has better results than the ARMA model with a decrease of 1.3 points while performing error prediction.	[58]
Corsica Island, France	AR and ARIMA	Historical global solar radiation	Daily global solar radiation	January 1998 to December 2007	RMSE, nRMSE, MAE and MBE	An ANN with extrinsic and intrinsic data has better performance in univariate ARMA models.	[59]
Seoul, South Korea	ARIMA and SARIMA	Historical solar radiation data	Daily and monthly solar radiation	1981–2017	R2 and RMSE	While the ARIMA model has good results for daily solar radiation, SARIMA has more accurate results for monthly solar radiation estimation.	[60]
Oran, Algeria	ARMANAR	Meteorological parameters	Hourly global horizontal solar radiation	2010–2012	RMSE and nRMSE	nRMSE is 0.2634 and 0.3241, respectively for NAR and ARMA.	[61]
Seoul, South Korea	SARIMA NARX	Historical global solar radiation	Global solar radiation	1981–2015	RMSE and R2	R2 and RMSE are 0.95 and 0.23 MJ/m2, respectively for NARX.	[62]
New Delhi, India	ARIMA	Meteorological parameters	Monthly solar radiation	July 01, 1983, to December 31, 2007	RMSE, MAPE, MAE, and R2	MAPE, R2, RMSE, and MAE results are 6.556, 0.9293, 0.3529, and 0.2659, respectively.	[63]
Las Vegas, United States	ARMA	Historical solar radiation	Hourly solar radiation	1995–2004	MBE and nRMSE	MBE and nRMSE are 0.133% and 11.76%, respectively.	[64]
Morocco	ARMA and ARIMA	Historical solar radiation	Daily global solar radiation	2018 (November to December) and 2019 (January to March)	MBE RMSE, and MAPE	ARIMA has a better performance compared to the ARMA model.	[65]
Two cities (Algiers and Ghardaia) in Algeria	AR and NAR	Historical solar radiation	Daily global solar radiation	1 January 2005 to 31 December 2006	NRMSE, S, R2, RMSE, MAE and MBE	NAR has a better performance compared to the AR model.	[66]
Ghardaia, Algeria	ARMA, NARX, and AR	Historical solar radiation	Hourly global solar radiation	May 2013 to October 2013	RMSE, NRMSE, MAPE, NMBE and R	NARX estimated the solar radiation data more accurately than other models.	[67]