Learning mobility user choice and demand models from public transport fare collection data
Frédéric Roulland, Luis Ulloa, Arturo Mondragon, Michael Niemaz, Guillaume Bouchard, Victor Ciriza
In this paper we present a new approach for public transit simulation which is based in machine learning techniques in order to model user choices and demand. We believe this will enable a new generation of much finer grained simulations for planning public transit infrastructure development making use of the richness of the information contained in the mass of data collected by modern transportation systems. We describe how we have implemented the system using public transport fare collection data, what is our current results and our next steps for the maturation of this technology.
ITS 21st World Congress, Detroit, 7-11 September, 2014
2014-003.pdf (200.56 kB)