SEARCH RESULT

In this study, regression learning methods such as linear regression, linear Support Vector Machines (SVM) and Gaussian SVM are used to estimate the wind speed on monthly time series. The wind speed data is consisted of ten-minute bars taken from the wind central in Zonguldak province in Turkey. In the pre-processing stage, Moving Average (MA), Weighted MA and Exponential MA filters are performed by using between 3 and 10 delay times on the wind speed data set. The data range is converted to the range [0, 1] in the normalization operation. Three different regression methods are used to estimate the wind speed. In the training phase of the models 10-fold crossvalidation method is used. The performance of the models is compared with statistical indicators such as Mean Absolute Error (MAE), Mean Squared Error (MSE) and Root Mean Squared Error (RMSE). The minimum estimation error value is determined for used models. It has been observed that Gaussian SVM model approach gives the least error to estimate wind speed with MA filters and delay steps when compared to other methods

1.st International Conference Energy Systems Engineering
ıcese'17

Seçkin Karasu Aytaç Altan Zehra Saraç Rıfat Hacıoğlu