Markov switching negative binomial models: an application to vehicle accident frequencies - Statistics > ApplicationsReportar como inadecuado




Markov switching negative binomial models: an application to vehicle accident frequencies - Statistics > Applications - Descarga este documento en PDF. Documentación en PDF para descargar gratis. Disponible también para leer online.

Abstract: In this paper, two-state Markov switching models are proposed to studyaccident frequencies. These models assume that there are two unobserved statesof roadway safety, and that roadway entities roadway segments can switchbetween these states over time. The states are distinct, in the sense that inthe different states accident frequencies are generated by separate countingprocesses by separate Poisson or negative binomial processes. To demonstratethe applicability of the approach presented herein, two-state Markov switchingnegative binomial models are estimated using five-year accident frequencies onIndiana interstate highway segments. Bayesian inference methods and MarkovChain Monte Carlo MCMC simulations are used for model estimation. Theestimated Markov switching models result in a superior statistical fit relativeto the standard single-state negative binomial model. It is found that themore frequent state is safer and it is correlated with better weatherconditions. The less frequent state is found to be less safe and to becorrelated with adverse weather conditions.



Autor: Nataliya V. Malyshkina, Fred L. Mannering, Andrew P. Tarko

Fuente: https://arxiv.org/







Documentos relacionados