Investigation of time series forecasting when series occur outliers

Forecasting is one of the most important concepts in time series. To forecast truly the model should be determined that identifies the data set best. However one or more outliers in the model affect the parameters of the model and forecasting. In the scope of this study, firstly, time series and the outliers in time series concepts are identified. The effect of outliers is investigated on the ARMA model parameters and forecasting. For this reason, data of TUIK are used to research outliers and forecasting.