Deteksi Stasioneritas Data Runtun Waktu Melalui Uji Akar-akar Unit

Abstract

One of the important types of data used in emperical analysis is time series data. In regressing a time series variable on another time series variable(s), one often obtains a very high R 2 even though there is no meaningful relationship between the two variables. Sometime we expect no relationship between the two variables, yet a regression of one on the other variable often shows a significant relationship. This situation exemplifies the problem of spurious regression. Spurious regression problem may arise from regressing a nonstationary time series variable on one or more nonstationary time series variable(s). Therefore, it considers important to be able to determine whether a time series data is stationary or not. A time series data is called stationary if it does not contain any unit roots. A test of stationary (or nonstationary) has become widely popular since the last two decades is the unit root test. This paper is meant to explain how to use unit root test in time series model especially for an autoregressive time series model. Key words: autoregressive time series, stationary, unit root