PERAMALAN PASANG SURUT AIR LAUT DI PULAU JAWA MENGGUNAKAN MODEL GENERALIZED SPACE TIME AUTOREGRESSIVE (GSTAR) (Studi Kasus : Ketinggian Pasang Surut Air Laut di Stasiun Pasang Surut Jakarta, Cirebon, Semarang dan Surabaya)

Chyntia Arum Widyastusti
Abdul Hoyyi
Rita Rahmawati

Jurnal Gaussian
Vol 5, No 4 (2016): Jurnal Gaussian

In daily life is often found time series data contains not only connection among  the events in previous times, but also has a relationship between one location to another. Data with time series and location linkage is called space-time data. Generalized Space Time Autoregressive (GSTAR) model is one of the commonest used to make model and forecast space-time data. The purposes of this research are to get the best GSTAR model and the forecasting results for the data ocean tide heights at four stations of Java island, those are Stations of Jakarta, Cirebon, Semarang and Surabaya. The best model obtained is GSTAR(1;1)-I(1) which is using cross correlation normalization weight because its residuals fulfill white noise assumption with the smallest value of MAPE and RMSE. The best GSTAR model explains that the elevation ocean tide data in Stations of Cirebon and Semarang is only influenced by the earlier times, and not influenced by other locations but can affect the height of the tide at other locations. As for the elevation ocean tide data stations of Jakarta and Surabaya are influence each other. Keywords: GSTAR, Space-Time, Ocean Tide, MAPE and RMSE.