01504237

Component

http://onlinepubs.trb.org...es/2011/RSS/1/Matkan,A.pdf

In recent years considerable studies have been carried out to investigate the relationships between crashes and related factors. Trip generator variables as the function of socio-economic and demographic characteristics and land use along with exposure variable, Vehicle Kilometers Travelled (VKT) might be appropriate variables to estimate the crashes at aggregated area-level. Generalized Linear Models (GLMs) as the most widely used global models employed in crash analyses, describe the relationships between the dependent and independent variables by estimating the fixed global coefficients. Since the crash occurrences are likely influenced by many spatial factors which vary over space, Geographically Weighted Poisson Regression (GWPR) is introduced to investigate the non-stationary effects in estimating the number of crashes per Traffic Analysis Zone (TAZ). In this paper, the GWPR was primarily employed on 253 TAZs in Mashhad, Iran, then Moran’s I indicator was performed on model residuals to examine the spatial autocorrelation. GWPR shows an improvement in model performance as indicated by goodness-of-fit measures. The results also indicate the non-stationary state in the relationships between the number of crashes and independent variables. The positive sign of coefficients for most of the TAZs shows that although the selected variables have a positive influence on the number of accidents, their magnitude varies considerably over the space. Moreover, testing the GWPR residuals conveys that the model has successfully removed the spatial autocorrelation among residuals.

Distribution, posting, or copying of this PDF is strictly prohibited without written permission of the Transportation Research Board of the National Academy of Sciences. Unless otherwise indicated, all materials in this PDF are copyrighted by the National Academy of Sciences. Copyright © National Academy of Sciences. All rights reserved

01501394

English

16p

2011

3rd International Conference on Road Safety and Simulation

Digital/other

Figures; Maps; References; Tables

Goodness of fit; High risk locations; Highway safety; Land use planning; Spatial analysis; Traffic crashes; Trip generation; Urban areas

Generalized linear mixed model; Geographically weighted regression

Mashhad (Iran)

Highways; Planning and Forecasting; Safety and Human Factors; I72: Traffic and Transport Planning; I80: Accident Studies

TRIS, TRB, ATRI

Jan 9 2014 4:38PM