Predicting Wind Speed in the Area of Baharka

Abstract

AbstractThe aim of research is to use the steps of analysis ond to clarify them through using Auto Regressive Integrated Moving Average (ARIMA). Also, it aims at examining the suitability of the model diagnosed in predicting. Time series of wind speed is used weekly with an average of 182 observations, from 1/1/2008 to 28/6/2011. These observations are analyzed via using (Box-Jenkins) models. Depending on the theoretical models and comparing them to time servies of wind speed, the model ARIMA (2,1,1) is chosen. However, the parameters of the adopted model were insignificant; therefore, MSE, AIC, AICc, BIC criteria are adopted in order to choose the best model.The result show that ARIMA (0,1,1) model is the most suitable one for it has the least values of the a formentioned four criteria. Suitability tests are used to ensure the suitability of ARIMA (0,1,1) model. Then, the latter is used to predict future time series. The researcher notes that this model gives similar values for the predicted peroids with trust peroids differences.