Citation

Tracz M, Nowakowska M. Transp. Traffic Theory 1996; 13: 261-276.

Copyright

(Copyright © 1996, Publisher varies)

DOI

unavailable

PMID

unavailable

Abstract

The paper presents a probabilistic model developed for the evaluation of the level of road sections safety. Cluster analysis was used to determine the sections of a road that should be examined in order to identify blackspots among them. To process this identification, Bayesian theory has been adopted. It uses the general information about accident features distribution over all the sections as prior knowledge, and a piece of information about the features distribution specific to a certain section as posterior knowledge. Each road section is then characterized by a certain weight value derived from the prior and the posterior knowledge. This weight is then used as a basic value in the identification process, in which the section is or is not classified as a blackspot one. The model was practically applied to accident data obtained from police reports. The results of this application are presented in the last part of the paper. (A) For the covering abstract see IRRD 886400.