Citation

Khaksar H, Almasi SA, Goharpoor AA. J. Transp. Res. (Tehran) 2022; 19(1): 45-58.

Copyright

(Copyright © 2022, Iran University of Science and technology, Transportation Research Institute)

DOI

10.22034/tri.2021.250276.2815

PMID

unavailable

Abstract

Identifying road segment at risk of accidents offers a special approach to safety professionals to better understand crash patterns and enhance road safety management. Conventional methods for identifying accident hotspots and crash patterns are not strong enough to take into account the spatial properties of crash data in the model. Traffic accidents with a spatial nature tend to be spatially dependent, Spatial models describe the predicted value of the crash pattern in space, which can be due to changes in the remarkable properties of the local environment Reflects crash densities better and provides a more realistic picture of crash distribution. In this study, all the main suburban axes of Hamedan province based on spatial accident data from 2017 to 2019 using kernel density distribution methods, geographical weighted regression, (GWR) geographical weighted Poisson regression(GWPR) have been studied. The results of the models show that the geographically weighted Poisson regression(GWPR) model has better results for predicting crash locations than other models.

Language: en