Pemodelan Generalized Space Time Autoregressive (GSTAR) Pada Data Inflasi di Kota Samarinda dan Kota Balikpapan

Abstract

One of the macroeconomic indicators used in the preparation of government’s  economicpolicy is inflation. Inflation is a data time series monthly that also is influenced by location effects. Generalized Space Time Autoregressive (GSTAR) is a time series methode that combines time and location effects. The case study is applied of GSTAR for forecasting inflation in two cities in East Kalimantan namely Samarinda and Balikpapan. This research aims to implement GSTAR model to gain forecasting model for inflation data in Samarinda city and Balikpapan city by using method of cross-correlation normalization. The resulting model is GSTAR model GSTAR (2,1) and GSTAR (3,1). The model obtained is not feasible to be used for forecasting, because it does not meet the white noise assumption.