SARIMA MODELLING APPROACH FOR RAILWAY PASSENGER FLOW FORECASTING

Transport

Miloš Milenković¹  Zoran Avramović¹  Nebojša Bojović¹  Vlastimil Melichar²  Libor Švadlenka² 

[1] University of Belgrade;University of Pardubice;

关键词: railway;    passenger service;    time series;    forecasting;    SARIMA;   

DOI  :  10.3846/16484142.2016.1139623

来源: DOAJ

【 摘 要 】

In this paper, railway passenger flows are analyzed and a suitable modeling method proposed. Based on historical data composed from monthly passenger counts realized on Serbian railway network it is concluded that the time series has a strong autocorrelation of seasonal characteristics. In order to deal with seasonal periodicity, Seasonal AutoRegressive Integrated Moving Average (SARIMA) method is applied for fitting and forecasting the time series that spans over the January 2004 – June 2014 periods. Experimental results show good prediction performances. Therefore, developed SARIMA model can be considered for forecasting of monthly passenger flows on Serbian railways.

【 授权许可】

Unknown