Pedestrian Crash Exposure Analysis Using Alternative Geographically Weighted Regression Models

Abstract

In order to develop a sustainable, safe, and dynamic transportation system, proper attention must be paid to the safety of pedestrians. The purpose of this study is to analyze the surrogate measures related to pedestrian crash exposure in urban roads, including the use of sociodemographic characteristics, land use, and geometric characteristics of the network. This study develops pedestrian exposure models using geographical spatial models including geographically weighted regression (GWR), geographically weighted Poisson regression (GWPR), and geographically weighted Gaussian regression (GWGR). In general, the results of the GWPR model show that the presence of a bus station, population density, type of residential use, average number of lanes, number of traffic control cameras, and sidewalk width are negatively associated with increasing the number of crashes. In this study, in order to identify traffic analysis zones (TAZ) based on the observed and predicted crash data, spatial distance-based methods using GWPR outputs have been used. This study shows the dispersion and density of pedestrian crashes without possessing the volume of pedestrians. Comparison of the performance of GWPR and Poisson models shows a significant spatial heterogeneity in the analysis.

Using Spatial Modeling for Identifying Determinants Influencing Crash Estimate: Case Study of Hamedan, Iran

“…The difference between the traditional generalized linear model (GLM) and GWPR models is that GLM has a constant parameter for each variable ( 6 , 10 ), whereas GWR Software has different spatial parameters for each variable. The parameter of a variable in the traditional model falls within the range of spatial model parameters of the same variable, representing an estimated parameter ( 45 , 46 ).…”

“…Finally, the models were categorized into clusters based on the presence or absence of spatial autocorrelation among the independent variables. Drawing from previous research ( 10 , 12 , 48 ), both global (nonspatial) and spatial models were employed to estimate crash frequency. Subsequently, we utilized Gaussian process regression and GWPR techniques, depending on the presence of spatial autocorrelation between the independent variables and spatial correlations between the independent and dependent variables.…”

“…Based on the suggestion of HSM, the specified overdispersion in crash data can be well predicted using the empirical Bayes method with the negative binomial model (NB) ( 8 , 9 ). The NB model appropriately adjusts overdispersion in the data and the Poisson model is used to predict the crash frequency ( 10 – 12 ). Additionally, researchers have stated that localized SPFs can better represent traffic-crash frequencies and associated factors (such as AADT and road segment length), which can be attributed to spatial dependence between crash observations ( 9 , 13–15 ).…”

The Influence of Demographic Factors on Urban Medical Insurance Expenditure under Big Data

“…First, the regression analysis [27][28][29][30][31][32][33][34][35] was performed on lny, lnx 1 , x 2 , x 3 , x 4 , the residual sequence of the equation, named E, was tested for unit root stationarity for E, and the results are shown in Figure 2.…”

Drivers' performance assessment approaching pedestrian crossings through the analysis of the speed and perceptive data recorded during on-field tests

“…A possible approach to research this topic is the passive approach, which is based on the analysis of crash database in order to identify of both crash-prone locations and factors affecting the phenomenon [ 7 , [21] , [22] , [23] , [24] ]. However, this approach fails to go into detail and does not investigate the crash environment or the road characteristics approaching the site of the crash [ 25 ] and often does not consider driving speed as a variable.…”