This paper explores how geometric characteristics, accesses, infrastructures and road environment influence crash prediction and safety performance of rural roads in the Dutch province of Gelderland. More than 500km two-way two-lane rural road segments in Gelderland were investigated and 4968 groups of data were collected with length unit of 100m. Safety performance of these roadway segments was evaluated and classified based on ten-year observed crash data. Mechanical language and Naive Bayes classifier were applied to the crash prediction model and a 10-fold cross-validation was conducted to evaluate the performance of the model under the guarantee that it was not over fitted. The results reveal how road conditions and environmental elements, especially access elements and object elements, contribute to road safety. It provides a new prospective to improve road safety performance.
Research on Crash Prediction and Safety Performance of Two-Way Two-Lane Rural Roads in Gelderland
15th COTA International Conference of Transportation Professionals ; 2015 ; Beijing, China
CICTP 2015 ; 3071-3080
13.07.2015
Aufsatz (Konferenz)
Elektronische Ressource
Englisch
Safety performance functions for crash severity on undivided rural roads
Online Contents | 2016
|Safety Improvements on Two-Lane Rural Roads
British Library Conference Proceedings | 2000
|Safety Improvements for Two-Lane Rural Roads
NTIS | 2000
|Crash Modification Functions for Passing Relief Lanes on Two-Lane Rural Roads
Transportation Research Record | 2020
|Crash reduction following installation of centerline rumble strips on rural two-lane roads
Online Contents | 2004
|