Forecasting Traffic Flow Conditions in an Urban Network: Comparison of Multivariate and Univariate Approaches

Abstract

Several univariate and multivariate models have been proposed for performing short-term forecasting of traffic flow. Two different univariate [historical average and ARIMA (autoregressive integrated moving average)] and two multivariate [VARMA (vector autoregressive moving average) and STARIMA (space–time ARIMA)] models are presented and discussed. A comparison of the forecasting performance of these four models is undertaken with data sets from 25 loop detectors located in major arterials in the city of Athens, Greece. The variable under study is the relative velocity, which is the traffic volume divided by the road occupancy. Although the specification of the network’s neighborhood structure for the STARIMA model was relatively simple and can be further refined, the results obtained indicate a comparable forecasting performance for the ARIMA, VARMA, and STARIMA models. The historical average model could not cope with the variability of the data sets at hand.

Gaussian Process Decentralized Data Fusion and Active Sensing for Spatiotemporal Traffic Modeling and Prediction in Mobility-on-Demand Systems

“…It is also not shown how these models can be generalized to work for arbitrary road network topologies and more complex correlation structure. Existing multivariate parametric traffic prediction models [41], [42] do not quantify uncertainty estimates of the predictions and impose rigid spatial locality assumptions that do not adapt to the true underlying correlation structure.…”

A visual analytics framework for spatio-temporal analysis and modelling

“…The framework also allows the analyst to control the degree of data abstraction and generalisation and achieve a suitable trade-off between the model quality and model complexity, i.e., the number of different statistical models that represent the entire spatio-temporal variation. Kamarianakis and Prastacos (2003) compared several methods for modelling spatio-temporal data and found that a global spatio-temporal model does not necessarily perform better than a set of local temporal models. One of the arguments in favour of a global model was the excessive computational time needed for building multiple local models in case of a very large dataset.…”

Limits of Predictability for Large-Scale Urban Vehicular Mobility

“…The related work in vehicular mobility prediction comes from the two main communities, i.e., the transportation system community [1], [2], [9], [22] and the vehicular networking community [13], [14]. From the first community, some works investigate the vehicular behavior prediction in terms of trajectories and routes [1], [6]- [9], [22].…”

“…of the resource of networks of roads and transportation systems, becomes a hot research topic that attracts broad interests [2]. On the other hand, newly emerged vehicular communication networks are seen as a key technology to help in relieving the traffic congestion and improving road safety, by building intelligent transportation systems [3]- [5].…”