Abstract
Semarang as the capital of the province of Central Java is a central transportation that has a high intensity and strategic activities. However, this area has a tidal disaster threat level is high enough. Tidal flood is a phenomenon where sea water entered the land area when the sea level has getting tides. In the future impact of tidal inundation in Semarang city is predicted to be greaterso that has needed the forecasting of high tide. The data pairs tend to experience seasonal monthly and contained outliers that may affect the suitability of the model so that Seasonal Autoregressive Integrated Moving Average (SARIMA) and outlier detection is used for forecasting method. For outlier detection, there are four types of outliers are additive outlier (AO), innovational outlier (IO), level shift (LS) and temporary change (TC). The study was conducted on the data of tide in Semarang period January 2004 - December 2012 based on the average high tide occurs when the maximum. The results of research showed that the model SARIMA with 7 outliers result predictions with high accuracy because it has a smaller AIC value is 649,1083 compared to the SARIMA models without outlier is 705,6404.
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have