Abstract

Wind speed data collection process faces several problems as failure of data observing devices. Therefore, windspeed data naturally contains missing values. Imputing these missing values using an effective method isimportant before performing time series analysis. The classical methods as linear, nearest neighbor, and statespace may not provide accurate imputations when the wind speed contains nonlinearity. In this study, the hybridartificial neural network (ANN) and autoregressive (AR) method is proposed for imputing the missing values.ANN is a nonlinear method that is capable of imputing the missing values in wind speed data with nonlinearcharacteristic. AR model is used for determining the structure of the input layer for the ANN. Listwise deletion isused before AR modeling to handle the missing values. A case study is carried out using daily Iraqi andMalaysian wind speed data. The proposed imputation method is compared with linear, nearest neighbor, andstate space methods. The comparison has shown that AR-ANN outperformed the classical methods. Inconclusion, the missing values in wind speed data with nonlinear characteristic can be imputed more accuratelyusing AR-ANN. Therefore, imputing the missing values using AR-ANN leads to more accurate performance oftime series modeling and analysis.

This work is licensed under a Creative Commons Attribution 4.0 License.