Short-Term Wind Speed Forecasting using Long Shot-Term Memory Algorithms

Accurate wind speed prediction is crucial for the operation and planning of wind energy generation. However, this task presents significant challenges due to the wind's unpredictable and unstable nature. Various machine learning algorithms have been proposed to address these challenges and enhance the reliability of wind speed forecasting. This paper proposes the development of machine learning algorithms, including support vector regression (SVR), extreme gradient boosting (XGBoost), random forest regression (RFR), and long short-term memory (LSTM) networks, to approach short-term wind speed forecasting. The researchers applied time series data with a 1-hour time interval from the NASA power project for the Kanyakumari site in the 2023 dataset to the input data models, utilizing wind speed values measured at heights of 10 m and 50 m. Different models' wind speed forecasting performance is evaluated using various accuracy measures, namely mean square error, root mean square error, mean absolute error, mean absolute percentile error, and coefficient of determination. The results indicate that the LSTM model demonstrates superior prediction accuracy and stability compared to the other models.