VALIDATION OF SHORT-TERM PREDICTION FOR ANNUAL AVERAGE DAILY TRAFFIC IN HONG KONG

The most up-to-date annual average daily traffic (AADT) flow is always required for transport model development and calibration. However, the current-year AADT data are not always available. This study presents time series models based on historical traffic counts and available current-year partial traffic counts to predict the daily traffic flows by days of the week and of months as well as the AADT for the whole current year in Hong Kong. In this paper, we have chosen 2 groups of core stations to develop the short-term traffic forecast model using Auto-Regressive Integration Moving Average (ARIMA) method. The time series model of each core station is calibrated and validated against each of the observed daily flows for all 7 days of the week. These models are then used to validate the AADT of the current and succeeding years. The results indicate that the time series models can be used effectively in predicting AADT in the short-term. The shortage of core stations affect the accuracy of the estimation results for a whole group.