<![CDATA[A model that combines both time and location factors in a multivariate time series called space time model. Generalized Space Time Autoregressive (GSTAR) is one of the space-time models that can be utilized in data forecasting. The GSTAR model is a development method of Space Time Autoregressive (STAR) that can be used for heterogeneous locations. The GSTAR model is used to forecast data in multiple locations at once. Generalized Least Squares (GLS) is one of the estimation methods that can be used in the GSTAR model. The GLS method is used on data that has residuals that are correlated across equations. This study applies GSTAR model to forecast farmer exchange rates of horticultural subsector in West Java, Yogyakarta, East Java, and Banten using GSTAR-GLS (11)I(1) model with uniform, invers distance, and cross-correlation normalization weights. The analysis result for model GSTAR-GLS (11)I(1) with three weighted methods shows that the best forecast result is using uniform weights with SMAPE for West Java, Yogyakarta, East Java, and Banten are 2.85%; 3.63%; and 1.92% or the forecasting result is highly accurate.]]>
Copyrights © 2026
