Analysis and forecast of Turkey unemployment rate

Abstract

This research develops techniques which are useful in forecasting single variable time series data. The techniques used in this study are moving averages (MA), Single Exponential Smoothing (SES), Adaptive Response Rate Exponential Smoothing (ARRES), Holt’s Linear and Holt-Winter’s Trend and Seasonality. For the purpose of this study, secondary data of Turkey Unemployment Rate covering the period 1996 up to 2015 was obtained from the Turkish Statistical Institute (TurkStat). From the result obtained, Adaptive Response Rate Exponential Smoothing (ARRES) was found to be the best method to forecast the Turkey Unemployment rate since it produces the lowest Mean Square Error (MSE) value which is 1.519.