Using artificial intelligence for wind speed prediction

Abstract

Accurate wind speed prediction is critical for renewable energy management, agriculture, and weather forecasting. This study investigates the use of machine learning techniques for predicting daily average wind speed using meteorological features, including rainfall, maximum and minimum temperature, grass minimum temperature, and atmospheric indicators. Two models; Linear Regression and Random Forest Regressor were implemented and evaluated on a dataset containing 6, 574 daily records from a meteorological station. Performance was assessed using Mean Absolute Error (MAE), Mean Squared Error (MSE), Root Mean Squared Error (RMSE), and R-squared. Results indicate that Random Forest achieved superior predictive accuracy (MAE = 3.71, RMSE = 4.69, R² = 0.71) compared to Linear Regression (MAE = 3.87, RMSE = 4.85, R² = 0.66). The findings demonstrate the potential of ensemble learning methods in improving wind speed forecasting accuracy.