Estimation of monthly average daily of the global solar radiation using the linear regression algorithm

Abstract

Abstract Solar radiation is the backbone for the existence of life on the earth. Its measurement is too expensive. Therefore, it has been great importance to propose an efficient method to use as a solar predictor based on other more readily available meteorological data. In this study, Linear regression method has been applied in two locations that have two different climates in Egypt, location A that has a coastal climate on the Mediterranean coast and location B that has a dry desert climate, using three weather vectors (minimum, maximum and average) of temperature values. The performance of the linear regression models provides better predictions for global solar radiation (GSRa) at different locations. The Root Mean Square Error (RMSE) = (2.7716 and 2.9392) MJ m−2 with Relative Root Mean Square Error (rRMSE) = (14.5962 and 14.0861) % in location A and B respectively. The accurate prediction results of the GSRa using this approach can be employed in the various purposes of the solar applications.