Predicting Solar Radiation in Manabi: A Machine Learning Approach

Solar radiation prediction has been studied within the academic context in several geographical locations to determine the variables that affect this phenomenon. However, there are no specific studies that have been carried out in the province of Manabi in Ecuador. To fill this gap in knowledge, this study develops a predictive model using machine learning algorithms to estimate solar radiation. A quantitative methodology is used with data collected from the TJ-JUNAYA weather station located in the Playa Prieta area of Manabi. The results indicate that the XGBoost algorithm is the most effective for predicting solar radiation based on RMSE, MSE, and R-2 metrics, with the UV index being the most influential variable. This study is important for two main reasons. First, it helps people in make better decisions by giving them tools to predict sunlight, which can make solar energy systems work better. Second, it uses weather details that were not used before to make better prediction models.