01504259

Component

http://onlinepubs.trb.org...SS/1/Aguero-Valverde,J.pdf

Spatial correlation has shown to be present in highway safety data, yet the distance at which sites should be considered correlated is largely unknown. The purpose of this research is to explore the effect of direct spatial correlation structures on crash frequency models at the road segment level and compare them to spatial conditional models. By using direct spatial correlation structures the “effective range” (i.e. the distance at which there is no lingering spatial correlation) is estimated. A Full Bayes hierarchical approach is used with direct spatial correlation effects for the spatial correlation terms as well as conditional autoregressive (CAR) spatial effects. The model of crash, traffic and roadway inventory data from Pennsylvania shows an average effective range of around 168 m (0.1 mi) which confirmed previous findings from conditional models that suggested that spatial correlation is presented in segments less than a mile apart. The direct spatial correlation model has a better goodness-of-fit than the random effects and the CAR model. In terms of posterior deviance the direct spatial correlation model performed similar to the other two models but in terms of the penalized goodness-of-fit measure Deviance Information Criteria it performed significantly better (4203, 4181, and 4109 respectively). Furthermore, the proportion of variation in the data explained by the spatial correlation term is almost the same for both spatial models (0.59 for the CAR model and 0.60 for the direct spatial correlation model). The standard deviations in coefficient estimates are slightly lower for the direct spatial correlation model compared to the random effects model but significantly lower than the CAR model

Distribution, posting, or copying of this PDF is strictly prohibited without written permission of the Transportation Research Board of the National Academy of Sciences. Unless otherwise indicated, all materials in this PDF are copyrighted by the National Academy of Sciences. Copyright © National Academy of Sciences. All rights reserved

01501394

English

13p

2011

3rd International Conference on Road Safety and Simulation

Digital/other

Figures; References; Tables

Bayes' theorem; Crash data; Crash rates; Goodness of fit; Spatial analysis; Traffic safety

Spatial correlation

Pennsylvania

Highways; Safety and Human Factors; I80: Accident Studies

TRIS, TRB, ATRI

Jan 9 2014 4:09PM