Original scientific paper
https://doi.org/10.7225/toms.v05.n02.001
Data Mining and Big Freight Transport Database Analysis and Forecasting Capabilities
Massimiliano Petri
; Dept. Civil and Industrial Engineering, University of Pisa, Pisa, Italy
Antonio Pratelli
; Dept. Civil and Industrial Engineering, University of Pisa, Pisa, Italy
Giovanni Fusco
; Centre National de la Recherche Scientifique, Université de Nice Sophia Antipolis, Nice, France
Abstract
Transport modeling in general and freight transport modeling in particular are becoming important tools for investigating the effects of investments and policies. Freight demand forecasting models are still in an experimentation and evolution stage. Nevertheless, some recent European projects, like Transtools or ETIS/ETIS Plus, have developed a unique modeling and data framework for freight forecast at large scale so to avoid data availability and modeling problems. Despite this, important projects using these modeling frameworks have provided very different results for the same forecasting areas and years, giving rise to serious doubts about the results quality, especially in relation to their cost and development time. Moreover, many of these models are purely deterministic. The project described
in this article tries to overcome the above-mentioned problems with a new easy-to-implement freight demand forecasting method based on Bayesian Networks using European official and available data. The method is applied to the Transport Market study of the Sixth European Rail Freight Corridor.
Keywords
Freight demand model; Bayesian networks; European freight corridor; Demand forecasting
Hrčak ID:
167821
URI
Publication date:
21.10.2016.
Visits: 1.685 *