Comprehensive Analysis and Forecasting of Wind Speed at Khavda Using Machine Learning

This study investigates wind speed prediction at Khavda, Gujarat, using machine learning models. Long Short-Term Memory (LSTM), Support Vector Regression (SVR), Regression, and Random Forest algorithms were employed for long-term forecasting. The dataset comprised 1,014 entries starting from the year 1940, which included the monthly aggregated wind speeds and directions. Model performance was evaluated using R-squared (R<sup xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">2</sup>), Root Mean Squared Error (RMSE), Mean Squared Error (MSE), and Mean Absolute Error (MAE). While all models demonstrated suitability for forecasting, the SVR model proved most efficient with an R<sup xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">2</sup>value of 0.997, MAE of 0.1841, MSE of 0.0576, and RMSE of 0.2336. This research contributes to enhancing wind power prediction accuracy at Khavda, crucial for efficient grid management and resource allocation in renewable energy systems. The findings support the ongoing shift towards sustainable energy solutions and climate change mitigation efforts.