Generalized space-time autoregressive with exogenous variable (GSTAR-X) method and its application Available to Purchase

Time series is a group of observation values obtained at different points in time with the same time interval. In some empirical studies, time-series data often has its complexity. CPI data is time-series data, so it can be modeled using the time-series analysis method. Time series is a series of observations arranged sequentially according to time in a certain period of time-series data, current observations are related to previous observations. Time and location-related (spatiotemporal) modeling for CPI forecasting can be used. GSTAR stands for generalized space-time autoregressive. to improve prediction accuracy, Exogenous variables were added to the GSTAR model to transform it into the GSTARX model. Eid al-Fitr, Eid al-Adha, Christmas, and New Year, which are the results of calendar fluctuations, as well as increasing fuel prices, are exogenous variables included in GSTARX models for CPI forecasting. The case study in GSTARX modeling is applied to CPI forecasting for five cities in Java, namely Yogyakarta, Semarang, Tegal, Surakarta, and Purwokerto. The purpose of this study was to obtain a suitable GSTARX model for forecasting the CPI for 5 cities in Java. The best model for GSTAR and GSTARX is the normalized weight of cross-correlation with MSE 12.9% and RMSE 19.35%