Forecasting Surabaya - Jakarta Train Passengers with SARIMA model

Abstract

PT. Kereta Api Indonesia (Persero) DAOP 8 Surabaya is a State Owned Enterprise which operates rail transport services covering passenger and freight transport especially in East Java Province. This study aims to deterMine a forecasting model that able to provide prediction number of train passengers in relation Surabaya-Jakarta so as to help PT. Kereta Api Indonesia (Persero) DAOP 8 Surabaya anticipate the increase of train passengers by using Seasonal ARIMA (SARIMA) models. The data used in this research is secondary data from PT. Kereta Api Indonesia (Persero) DAOP 8 Surabaya on January 2012 until July 2017. Data collected then processed and analyzed using SARIMA method. Various SARIMA models were assessed and the best one was selected. Then, the models was evaluated based on normality of the residuals distribution, correspondence between the fitted and real amounts, and calculation of Sum Squared Residual. The results shown that the best time series model based on Minimum Sum Squared Residual (SSR) is SARIMA Model (0,1,1)(1,10)12. Based on analysis using SARIMA model (0,1,1)(1,10)12 obtained the forecasting of Surabaya - Jakarta train passengers in January 2018 until July 2018 ranged from 119,495 - 161,685. Higher volume of passengers flows in July 2018 resulted by school and college holidays. Seeing the increasing trend of passengers during peak season especially in July, it is expected that PT KAI Daop 8 Surabaya will improve performance so that the increasing number of railway passengers can be maintained.