Comparison of Multivariate Poisson lognormal spatial and temporal crash models to identify hot spots of intersections based on crash types

Cheng, Wen; Gill, Gurdiljot Singh; Dasu, Ravi; Xie, Meiquan; Jia, Xudong; Zhou, Jiao

doi:10.1016/j.aap.2016.11.022

Cited by 56 publications

(15 citation statements)

References 39 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Several studies recognizing the importance of unobserved heterogeneity have developed multivariate approaches that account for the potential dependency across count variables. The various model structures developed from multivariate models include multivariate Poisson regression model (Ye et al, 2009), multivariate Poisson lognormal model (Serhiyenko et al, 2016), multinomialgeneralized Poisson model (Chiou and Fu, 2013), multivariate Poisson gamma mixture count model (Mothafer et al, 2016), multivariate Poisson lognormal spatial and temporal model (Aguero-Valverde et al, 2016;Cheng et al, 2017), Integrated Nested Laplace Approximation Multivariate Poisson Lognormal model (Wang et al, 2017), Bayesian latent class flexible mixture multivariate model (Heydari et al, 2017) and multivariate random-parameters zeroinflated negative binomial model (Anastasopoulos, 2016).…”

Section: Earlier Researchmentioning

confidence: 99%

A joint econometric approach for modeling crash counts by collision type

Bhowmik

Yasmin

Eluru

2018

Analytic Methods in Accident Research

View full text Add to dashboard Cite

In recent years, there is growing recognition that common unobserved factors that influence crash frequency by one attribute level are also likely to influence crash frequency by other attribute levels. The most common approach employed to address the potential unobserved heterogeneity in safety literature is the development of multivariate crash frequency models. The current study proposes an alternative joint econometric framework to accommodate for the presence of unobserved heterogeneityreferred to as joint negative binomial-multinomial logit fractional split (NB-MNLFS) model. Furthermore, the study undertakes a first of its kind comparison exercise between the most commonly used multivariate model (multivariate random parameter negative binomial model) and the proposed joint approach by generating an equivalent log-likelihood measure. The empirical analysis is based on the zonal level crash count data for different collision types from the state of Florida for the year 2015. The model results highlight the presence of common unobserved effects affecting the two components of the joint model as well as the presence of parameter heterogeneity. The equivalent log-likelihood and goodness of fit measures clearly highlight the superiority of the proposed joint model over the commonly used multivariate approach. ` iv ACKNOWLEDGMENT I would like to convey my heartiest gratitude to my honorable supervisor Dr. Naveen Eluru for his excellent supervision and being a constant support in this thesis. I would also like to gratefully acknowledge Signal Four Analytics (S4A) and Florida Department of Transportation (FDOT) for providing access to Florida crash and geospatial data. I would like to convey my heartiest gratitude to Dr. Shamsunnahar Yasmin who helped me a lot throughout the journey in University of Central Florida and guide me through the thesis.

show abstract

Section: Earlier Researchmentioning

confidence: 99%

A joint econometric approach for modeling crash counts by collision type

Bhowmik

Yasmin

Eluru

2018

Analytic Methods in Accident Research

View full text Add to dashboard Cite

show abstract

“…Crash modeling includes severity modeling [16][17][18] and crash frequency modeling [19,20]. Spatial crash modeling belongs to crash frequency modeling, which contains multiple model structures [21][22][23][24][25][26], including Poisson lognormal model, negative binomial spatial model, Poisson lognormal spatial model, geographic weighted Poisson regression model, and Bayesian spatial varying-coefficient model. Since the purpose of this study is to examine the effect of traffic flow characteristics, two Bayesian lognormal models with different CAR priors were applied, since they have been widely applied in many different research fields such as epidemiology.…”

Section: Spatial Model Configurationmentioning

confidence: 99%

Developing a New Spatial Unit for Macroscopic Safety Evaluation Based on Traffic Density Homogeneity

Wang

Liu

2020

Journal of Advanced Transportation

View full text Add to dashboard Cite

Macrolevel crash modeling has been extensively applied to investigate the safety effects of demographic, socioeconomic, and land use factors, in order to add safety knowledge into traffic planning and policy-making. In recent years, with the increasing attention to regional traffic management and control, the safety effects of macrolevel traffic flow parameters may also be of interest, in order to provide useful safety knowledge for regional traffic operation. In this paper, a new spatial unit was developed using a recursive half-cut partitioning procedure based on a normalized cut (NC) minimization method and traffic density homogeneity. Two Bayesian lognormal models with different conditional autoregressive (CAR) priors were applied to examine the safety effects of traffic flow characteristics at the NC level. It was found that safety effects of traffic flow exist at such macrolevel, indicating the necessity of considering safety for regional traffic control and management. Furthermore, traffic flow effects were also examined for another two spatial units: Traffic Analysis Zone (TAZ) and Census Tract (CT). It was found that ecological fallacy and atomic fallacy could exist without considering traffic flow parameters at those planning-based levels. In general, safety needs to be considered for regional traffic operation and the effects of traffic flow need to be considered for spatial crash modeling at various spatial levels.

show abstract

“…Still at roundabouts road crashes occur because of various factors including the driver, vehicle, roadway and environment. There are several road safety researches conducted with the aim to: estimate the effects of roundabout geometric features [1][2][3][4][5], identify the potential locations for safety improvement [6][7][8][9][10][11], improve the vehicular operation [12][13][14], account the driver behaviour [15,16], and examine the before-and after scenarios [17][18][19]. Modifying geometric feature of the roundabouts has been found to reduce the number of crashes, in particular these treatments can be design to support a particular type of potential crashes.…”

Section: Introductionmentioning

confidence: 99%

Investigating the Possibilities of Specific Crash Type at Roundabouts

Somasundaraswaran¹,

Richardson²

2019

JTTE

View full text Add to dashboard Cite

A comprehensive analysis of road crashes is extremely supportive in identifying existing deficiencies and in proposing new the design strategies. There are several research studies carried out for improving safety at roundabouts with the aim to altering conflict types, reducing the drivers' speeds and crash severity, and highlighting a considerable differences in the safety performance of individual or group of roundabouts. The main purpose in this study was to explain the correlation between the types of crashes to the geometry of the roundabouts based on data collected from ten top poorly performing roundabouts in Toowoomba, Australia. Preliminary analysis revealed the following as top three types as prominent at these identified roundabouts such as angle crashes (74%), hit-object crashes (13%), and Rear-end crashes (9%). Excess proportion of specific crash method has been used as methodological tool to estimate the potential possibilities for a specific crash type exceeding threshold at each roundabout. Results show that eight out of ten roundabouts were regarded for a particular type of crash but only two roundabouts were viewed for two types of crashes such as angle and hit-object.

show abstract

Comparison of Multivariate Poisson lognormal spatial and temporal crash models to identify hot spots of intersections based on crash types

Cited by 56 publications

References 39 publications

A joint econometric approach for modeling crash counts by collision type

A joint econometric approach for modeling crash counts by collision type

Developing a New Spatial Unit for Macroscopic Safety Evaluation Based on Traffic Density Homogeneity

Investigating the Possibilities of Specific Crash Type at Roundabouts

Contact Info

Product

Resources

About