Multivariate forecasting of road traffic flows in the presence of heteroscedasticity and measurement errors


Address for correspondence: Osvaldo Anacleto, Department of Mathematics and Statistics, The Open University, Milton Keynes, MK7 6AA, UK.


Summary.  Linear multiregression dynamic models, which combine a graphical representation of a multivariate time series with a state space model, have been shown to be a promising class of models for forecasting traffic flow data. Analysis of flows at a busy motorway intersection near Manchester, UK, highlights two important modelling issues: accommodating different levels of traffic variability depending on the time of day and accommodating measurement errors due to data collection errors. This paper extends linear multiregression dynamic models to address these issues. Additionally, the paper investigates how close the approximate forecast limits that are usually used with the linear multiregression dynamic model are to the true, but not so readily available, forecast limits.