Abstract
The estimation of global horizontal irradiance (GHI) is crucial for assessing solar energy potential, especially for investment purposes in specific regions. This study employs two feature selection techniques such as recursive feature elimination (RFE) and least absolute shrinkage and selection operator (LASSO) to identify key variables from two datasets, which are then used to train four machine learning (ML) models such as Decision Tree (DT), Random Forest (RF), Extreme Gradient Boost (XGB), and Extra Trees (ET) regressors. The performance of these models is evaluated using three statistical metrics such as mean absolute error (MAE), root mean squared error (RMSE), and R-squared (R 2). The results show that the ET regressor, when combined with LASSO, achieves the best predictive performance, with an MAE of 1.36 W/m² and an RMSE of 2.46 W/m2. The study further employs Shapley Additive Explanations (SHAP) to interpret the model, revealing that parameters like Diffuse Horizontal Irradiance, Solar Zenith Angle, and Direct Normal Irradiance significantly impact GHI prediction accuracy. The combination of feature selection, advanced ML models, and SHAP analysis offers a comprehensive and transparent framework for solar energy resource assessment, addressing the need for accuracy and interpretability in GHI estimation.
Published Version
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call