Estimating Long-Term Delay Risk with Generalized Linear Models

This paper presents an original methodology to estimate delay risk a few days before operations with generalized linear models. These models represent a given variable with any distribution from the exponential family, allowing to compute for any subject its own probability distribution according to its features. This methodology is applied on small delays (less than 20 minutes) of high-speed trains arriving at a major french station. Several distributions are tested to fit delay data and three scenarios are evaluated: a single GLM with a negative binomial distribution and two two-part models using both a logistic regression as first part to compute the probability of arriving on time, and a second part using a negative binomial or a lognormal distribution to obtain the probabilities associated with positive delay values. This paper also proposes a validation methodology to assess the quality of these probabilistic prediction based on two aspects: calibration and discrimination.

[1]  Otto Anker Nielsen,et al.  Application of Data Clustering to Railway Delay Pattern Recognition , 2018 .

[2]  Antony Stathopoulos,et al.  Methodology for processing archived ITS data for reliability analysis in urban networks , 2006 .

[3]  Kornchawal Chaipah,et al.  Improving arrival time prediction of Thailand's passenger trains using historical travel times , 2014, 2014 11th International Joint Conference on Computer Science and Software Engineering (JCSSE).

[4]  Alex Landex,et al.  Methods to estimate railway capacity and passenger delays , 2008 .

[5]  Davide Anguita,et al.  Dynamic Delay Predictions for Large-Scale Railway Networks: Deep and Shallow Extreme Learning Machines Tuned via Thresholdout , 2017, IEEE Transactions on Systems, Man, and Cybernetics: Systems.

[6]  Jianxin Yuan,et al.  Stochastic modelling of train delays and delay propagation in stations , 2006 .

[7]  Jan Peters,et al.  Prediction of Delays in Public Transportation using Neural Networks , 2005, International Conference on Computational Intelligence for Modelling, Control and Automation and International Conference on Intelligent Agents, Web Technologies and Internet Commerce (CIMCA-IAWTIC'06).

[8]  Paul Schonfeld,et al.  Analyzing passenger train arrival delays with support vector regression , 2015 .

[9]  Nils O.E. Olsson,et al.  Punctuality problems from the perspective of timetable planners in Sweden , 2017 .

[10]  D. Hosmer,et al.  Goodness of fit tests for the multiple logistic regression model , 1980 .

[11]  Rob M.P. Goverde,et al.  Delay distributions in railway stations , 2001 .

[12]  Ralph B. D'Agostino,et al.  Evaluation of the Performance of Survival Analysis Models: Discrimination and Calibration Measures , 2003, Advances in Survival Analysis.

[13]  Stanley Lemeshow,et al.  Goodness-of-Fit Testing for the Logistic Regression Model when the Estimated Probabilities are Small , 1988 .

[14]  Miguel A. Salido,et al.  An Assessment of Railway Capacity , 2008 .

[15]  R. Rigby,et al.  Generalized Additive Models for Location Scale and Shape (GAMLSS) in R , 2007 .

[16]  Pavle Kecman,et al.  Predictive modelling of running and dwell times in railway traffic , 2015, Public Transp..

[17]  Malachy Carey,et al.  Testing schedule performance and reliability for train stations , 2000, J. Oper. Res. Soc..

[18]  Alan Agresti,et al.  Modeling Nonnegative Data with Clumping at Zero: A Survey , 2002 .

[19]  David W. Hosmer,et al.  Applied Logistic Regression , 1991 .

[20]  Stanley Lemeshow,et al.  Standardizing the power of the Hosmer–Lemeshow goodness of fit test in large data sets , 2013, Statistics in medicine.

[21]  A. Agresti.  Foundations of Linear and Generalized Linear Models , 2015 .