Machine Learning Techniques Based Predictive Modeling for Wind Speed Forecasting

This research analyzes the efficacy of several Machine Learning (ML) methods in forecasting wind speed. The research specifically examines and contrasts the efficacy of NN (Neural Network), SVM (Support Vector Machine), GPR (Gaussian Process Regression), and DT (Decision Tree) models. The research utilizes historical wind speed data from the Dehradun area of India (30°19′27.95″N, 78°2′2.11″E). The investigation involves comparing the accuracy of these approaches and evaluating their computational efficiency and scalability. The results highlight the superior performance of the GPR technique, which demonstrates lower errors in wind speed prediction. Specifically, GPR achieved a RMSE (root mean square error) of 0.174776173, a MSE (mean square error) of 0.030546711, a MAE (mean absolute error) of 0.106447822, and a regression value (R2) of 0.964572701. The findings of this study have significant implications for environmental monitoring and renewable energy (RE) planning, as they provide insights into the most effective ML algorithms for accurate and efficient wind speed predictions.