The Best Spatial Weight Matrix Order of Radius for GSTARIMA

Some methods of statistical methods for modelling and forecasting, such as generalized space time autoregressive integrated moving average (GSTARIMA). The GSTARIMA model expresses each observation at time and space as a weighted linear combination of space and time. One of the most important in GSTARIMA model is the spatial weight matrix. The aim of this study is to choose the best spatial weight matrix order of radius. The best model GSTARIMA to the relative the Nilaparvata lugens Stäl attacks in Aceh province, North Sumatera province, West Sumatera province, Riau province, Jambi province, Bengkulu province, South Sumatera province, and Lampung province is GSTARIMA (1, 0, 0)3.