Vol. 127, No. 1 * Pages 1–142 * January - March 2023

Wind speed forecasting has found economic significance as it can increase operational efficiency. In this regard, an accurate forecast of wind speed is crucial in the application of wind resources. This study is intended to incorporate independent and output variables as the input of support vector regression (SVR) to forecast wind speed of Zanjan and Ahvaz stations in Iran. The independent variables were minimum, maximum, and mean temperatures, relative humidity, precipitation, average visibility, and dew point temperature. The incorporation of independent and output variables were conducted with principal component analysis (PCA) and differential weighting scheme (DWS), respectively. DWS combined the forecasts of linear regression, SVR, and group method of data handling (GMDH) in which the SVR showed the best The forecast of DWS outperformed the other three mentioned models. The incorporation of DWS and PCA (DWS-PCA) improved the forecasts and the capability of DWS-PCA as a novel method was significant in terms of forecast stability. The novel method can be a robust approach for wind speed forecasting in some subjects such as renewable energy, and meteorological decisions.