Research on the Prediction of Metropolitan Traffic Density in China using ARIMA Model

Abstract

This investigation targets to research on construction of prediction model of traffic density and discuss suitable methods to decrease heavy traffic. The data is obtained from the public website of Government of Shenzhen which provides updates of traffic monitoring, time sequence analysis is put in practical manipulation, and it is verified that traffic undergoes regular alternations responsible for specific time periods every day. Data sets of traffic density of routes are analysed using ACF and PACF testing procedures and ARIMA, to obtain an output of the prediction model of tendency of traffic congestion of related area. The prediction model does not have apparent invalidity in statistical perspectives, and illustrate approximate results about probable time portions that produces proliferating heavy-traffic phenomena. Relevant analysis does encounter limitation that the model may become unpredictable to sudden situations like temporary route construction, and unpleasant weather conditions which may result in chaotic circumstances to fluctuation of traffic for citizens.