Forecasting dynamic public transport Origin-Destination matrices with long-Short term Memory recurrent neural networks

A considerable number of studies have been undertaken on using smart card data to analyse urban mobility. Most of these studies aim to identify recurrent passenger habits, reveal mobility patterns, reconstruct and predict passenger flows, etc. Forecasting mobility demand is a central problem for public transport authorities and operators alike. It is the first step to efficient allocation and optimisation of available resources. This paper explores an innovative approach to forecasting dynamic Origin-Destination (OD) matrices in a subway network using long Short-term Memory (LSTM) recurrent neural networks. A comparison with traditional approaches, such as calendar methodology or Vector Autoregression is conducted on a real smart card dataset issued from the public transport network of Rennes Métropole, France. The obtained results show that reliable short-term prediction (over a 15 minutes time horizon) of OD pairs can be achieved with the proposed approach. We also experiment with the effect of taking into account additional data about OD matrices of nearby transport systems (buses in this case) on the prediction accuracy.

[1]  Chris Smith,et al.  Avoiding the crowds: understanding Tube station congestion patterns from trip data , 2012, UrbComp '12.

[2]  Mu-Chen Chen,et al.  Forecasting the short-term metro passenger flow with empirical mode decomposition and neural networks , 2012 .

[3]  Lutz Prechelt,et al.  Early Stopping - But When? , 2012, Neural Networks: Tricks of the Trade.

[4]  Lei Zhang,et al.  Short-term forecasting of high-speed rail demand: A hybrid approach combining ensemble empirical mode decomposition and gray support vector machine with real-world applications in China , 2014 .

[5]  Lei Yu,et al.  Short-term forecasting of transit route OD matrix with smart card data , 2011, 2011 14th International IEEE Conference on Intelligent Transportation Systems (ITSC).

[6]  Leo G. Kroon,et al.  Dynamic Forecast Model of Time Dependent Passenger Flows for Disruption Management , 2012 .

[7]  Bruno Agard,et al.  MINING PUBLIC TRANSPORT USER BEHAVIOUR FROM SMART CARD DATA , 2006 .

[8]  Licia Capra,et al.  Mining Public Transport Usage for Personalised Intelligent Transport Systems , 2010, 2010 IEEE International Conference on Data Mining.

[9]  Nigel H. M. Wilson,et al.  Potential Uses of Transit Smart Card Registration and Transaction Data to Improve Transit Planning , 2006 .

[10]  Yoshua Bengio,et al.  Learning long-term dependencies with gradient descent is difficult , 1994, IEEE Trans. Neural Networks.

[11]  Etienne Côme,et al.  A mixture model clustering approach for temporal passenger pattern characterization in public transport , 2015, 2015 IEEE International Conference on Data Science and Advanced Analytics (DSAA).

[12]  Wei Wang,et al.  Bus Passenger Origin-Destination Estimation and Related Analyses , 2011 .

[13]  Adam Rahbee,et al.  Origin and Destination Estimation in New York City with Automated Fare System Data , 2002 .

[14]  Feng Qian,et al.  A Recurrent Neural Network approach to traffic matrix tracking using partial measurements , 2008, 2008 3rd IEEE Conference on Industrial Electronics and Applications.

[15]  Martin Trépanier,et al.  Individual Trip Destination Estimation in a Transit Smart Card Automated Fare Collection System , 2007, J. Intell. Transp. Syst..

[16]  Ying Chen,et al.  Data Mining and Pattern Matching for Dynamic Origin–Destination Demand Estimation , 2015 .

[17]  J. van Leeuwen,et al.  Neural Networks: Tricks of the Trade , 2002, Lecture Notes in Computer Science.

[18]  Xiaolei Ma,et al.  Mining smart card data for transit riders’ travel patterns , 2013 .

[19]  Jimmy Ba,et al.  Adam: A Method for Stochastic Optimization , 2014, ICLR.

[20]  Jürgen Schmidhuber,et al.  Long Short-Term Memory , 1997, Neural Computation.

[21]  Patrick Gallinari,et al.  Mining Ticketing Logs for Usage Characterization with Nonnegative Matrix Factorization , 2014, MSM/MUSE/SenseML.

[22]  Eleni I. Vlahogianni,et al.  Optimized and meta-optimized neural networks for short-term traffic flow prediction: A genetic approach , 2005 .

[23]  Moshe E. Ben-Akiva,et al.  Estimation and Prediction of Time-Dependent Origin-Destination Flows with a Stochastic Mapping to Path Flows and Link Flows , 2002, Transp. Sci..

[24]  E. Côme,et al.  Understanding Passenger Patterns in Public Transit Through Smart Card and Socioeconomic Data: A case study in Rennes, France , 2014 .

[25]  Helmut Lütkepohl,et al.  Vector Autoregressive Models , 2011, International Encyclopedia of Statistical Science.

[26]  Alex Cui.  Bus passenger Origin-Destination Matrix estimation using Automated Data Collection systems , 2006 .

[27]  Catherine Morency,et al.  Smart card data use in public transit: A literature review , 2011 .

[28]  Nicholas Jing Yuan,et al.  Reconstructing individual mobility from smart card transactions: a collaborative space alignment approach , 2014, Knowledge and Information Systems.

[29]  Ka Kee Alfred Chu,et al.  Enriching Archived Smart Card Transaction Data for Transit Demand Modeling , 2008 .

[30]  Gerd Kortuem,et al.  Catch me if you can: Predicting mobility patterns of public transport users , 2014, 17th International IEEE Conference on Intelligent Transportation Systems (ITSC).

[31]  Peter White,et al.  What role for smart-card data from bus systems? , 2004 .

[32]  Pascal Vincent,et al.  Artificial Neural Networks Applied to Taxi Destination Prediction , 2015, DC@PKDD/ECML.

[33]  Jinhua Zhao,et al.  Estimating a Rail Passenger Trip Origin‐Destination Matrix Using Automatic Data Collection Systems , 2007, Comput. Aided Civ. Infrastructure Eng..