Model Generalized Space Time Autoregressive Integrated Moving Average
(Studi Kasus: Inflasi Kota Surabaya, Malang dan Kediri)
Abstract
Generalized Space Time Autoregressive Integrated Moving Average (GSTARIMA) is the models used to model the time series data containing spatial dependence between its locations (space time). GSTARIMA models can be used in economics data including export-import, exchange rate, production and inflation. Inflation is goods prices raising process in general. The inflation development in East Java is monitored through economic development in some big cities, such as Surabaya, Malang, and Kediri. Inflation is not only influenced by the previous times, but also has a correlation between one cities to other cities. The purpose of this research are to get the best GSTARIMA model and the forecasting results of inflation data in Surabaya, Malang and Kediri. Parameter estimation is performed using the ordinary least squares estimator (OLS) with the inverse distance weighting and cross correlation. The GSTARIMA(2,0,0)1 is the best inflation model of Surabaya, Malang, and Kediri with inverse distance weighting. In prediction, the result of those three cities inflation in 2017 is generally decreased trended. This can be seen from the inflation value which is around the value of 0,3 to almost close to the zero value.