New Achievement for Prediction of Highway Accidents

Authors

  • Gholamali Shafabakhsh Semnan University
  • Yousef Sajed Semnan University

DOI:

https://doi.org/10.4186/ej.2015.19.1.139

Keywords:

Highway Accidents, Regression Models, Section's length, Curvature, longitudinal slope and Pick hour volume.

Abstract

Most research has been carried out about crash modeling but little attention to the urban highway. The candidate set of explanatory parameters were: traffic flow parameters, geometric infrastructure characteristics and pavement conditions. Statistical analysis is done by SPSS on the basis of nonlinear regression modeling and during the analysis, principal components are identified to assist the principal component analysis method and more important variables recognized that could exhibit best description of crash occurrence on the basis of available logics. Results indicate that the number of accidents per year increase with: length, pick hour volume and longitudinal slope whereas decreases with radius. Presented models show that crash occurrence is increased with the increase in each of section's length, pick hour volume and longitudinal slope variables whereas it is increased with the decrease of curvature. The remarkable result in this study was the effect of longitudinal slope variable on crash occurrence.

Downloads

Download data is not yet available.

Author Biographies

Gholamali Shafabakhsh

Faculty of Civil Engineering, Semnan University, University Sq., P.O. Box 35196-45399, Semnan, Iran

Yousef Sajed

Faculty of Civil Engineering, Semnan University, University Sq., P.O. Box 35196-45399, Semnan, Iran

Published

Vol 19 No 1, Jan 30, 2015

How to Cite

[1]
G. Shafabakhsh and Y. Sajed, “New Achievement for Prediction of Highway Accidents”, Eng. J., vol. 19, no. 1, pp. 139-151, Jan. 2015.