Abstract
Prices and pollution are rising in tandem with the daily increase in demand for crude oil, which is
detrimental to the environment. In order to save the environment and provide humans a better energy source,
greater efforts are being made to generate clean energy. New problems and difficulties have surfaced as the
renewable energy industry expands. Energy transmission and storage systems have difficulties due to the
unpredictability of electricity output from wind turbines, solar power plants, and dams. Machine learning
methods have been applied to predict energy fluctuations in order to create a more dependable and efficient
energy system. One machine learning application in that field is forecasting the sun's Global Horizontal
Irradiance (GHI). In order to achieve long-term GHI forecasting, this article uses a variety of machine learning
techniques, including linear regression and long-short term memory (LSTM) algorithms. Furthermore, the
method of predicting the GHI irradiance by using various levels of linear regressors to get the optimal regressor
level that offers the lowest error for the testing set based on cross-validation is where this paper's relevance lies.
The results demonstrated that the regressor approach offers a lower error for a shorter computational time when
compared to a single vanilla LSTM system.