Forecasting Number of Train Passengers Using Time Series Regression Integrated Calendar Variation and Covid 19 Intervention

Mega Silfiani (Institut Teknologi Kalimantan)
Farida Nur Hayati (Institut Teknologi Kalimantan)

Jurnal Matematika Sains dan Teknologi
Vol. 25 No. 1 (2024)

The purpose of this study is to obtain a forecasting model for the number of train passengers using time series regression integrated with variations in the Islamic calendar and the effects of COVID 19. This study uses the number of train passengers in Jabodetabek, Java (Non-Jabodetabek), and Sumatra from January 2006 to December 2022 as the data source. Time series regression with variations of the Islamic calendar and the effects of COVID 19 for Jabodetabek, Java (non-Jabodetabek), and Sumatra has an RMSE value for each category of 7657,821; 2453.827 and 275.901. In general, the number of train passengers for all categories (Jabodetabek, Java, Sumatra) has a seasonality. In Jabodetabek and Sumatra, Eid al-Fitr has a big impact on the number of train passengers. Meanwhile, one month before Eid al-Fitr has a big impact on the number of train passengers in Java (Non Jabodetabek). In addition, the impact of COVID 19 significantly affected the number of train passengers for all categories.