Machine Learning-Based Wind Speed Prediction: A Study on Gradient Boosting Regressor Algorithm

This paper introduces an innovative machine learning model designed for predicting wind speed by harnessing data collected through the acquisition chain deployed at Ensit's SIME lab. The research encompasses a comprehensive dataset spanning from November 2022 to November 2023. The model's training phase, employing Gradient Boosting Regressor, utilized one year of this data, while the subsequent year was dedicated to testing and validation. Our findings reveal that the proposed methodology surpasses alternative models in forecasting wind speed over both short (one hour) and longer (one week) time horizons. The model exhibits a remarkable R2 score of 0.92, indicative of its high predictive accuracy, complemented by a minimal mean square percentage error (MSE) of 0.02. These results underscore the efficacy and reliability of our approach in enhancing wind speed forecasting precision, with potential implications for various applications in renewable energy and meteorological domains.