Solar Irradiance Prediction using meteorological data by ensemble models

High power quality electrical power grids can be deployed after accurate solar irradiance prediction at any particular location so as to minimize the physical expenses. Number of machine learning models can be applied to endogenous and exogenous datasets. Gathering historical irradiance data(endogenous) is complex task whereas meteorological data(exogenous) is easily and widely available on Internet. This paper proposed ensemble learning models based solar irradiance prediction scheme for NASA generated meteorological data and implemented experimental results in terms of RMSE,Accuracy, MAE and MSE to compare 2 ensemble classifiers and regressors models for short-term prediction. The results demonstrated that the proposed models are more accurate over existing machine learning and neural network algorithms for solar irradiance prediction.