Macroscopic modeling of pedestrian and bicycle crashes: A cross-comparison of estimation methods
Review articleOpen access
2016/08/01 Full-length article DOI: 10.1016/j.aap.2016.05.001
Journal: Accident Analysis & Prevention
Abstract:
Highlights•Cross-comparison of estimation methods: non-spatial negative binomial, random parameter negative binomial, and Poisson-Gamma-CAR.•Inclusion of a comprehensive set of exposure and risk proxies for pedestrians and bicycles.•The random parameter negative binomial model performed better than other selected methods.•Percentage of commuters cycling or walking to work, percentage of households without motor vehicles, and mixed land use have a significant and positive correlation with the number of crashes.
Request full text