The Application Of X-12-Arima On Moving Seasonal Time Series Data: A Comparative Study Approach

Abstract

Time series data that portray economic pattern are usually complicated by factors influencing the value of data, which unavoidably obscure accurate interpretation. The most common and important source of disturbance is seasonal variation. In this paper, two methods often applied on seasonal time series data are compared. The first method is the classic decomposition and the second is based on regARIMA model, namely X-12-ARIMA. The discussion is illustrated through the application of both methods to the series of cement consumption in Indonesia. It is found that X-12-ARIMA captures the moving seasonal pattern, which is hard adjusted by the classic decomposition method. In term of technicality X-12-ARIMA is also easy to implement. Hence, to this type of time series data, X-12-ARIMA is recommended.