A Space–Time Multivariate Bayesian Model to Analyse Road Traffic Accidents by Severity

Boulieri, Areti; Liverani, Silvia; Hoogh, Kees de; Blangiardo, Marta

doi:10.1111/rssa.12178

Cited by 42 publications

(56 citation statements)

References 69 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…Crash occurrences across severity outcomes are not independent, so the use of a multivariate model when different crash severities need to be analyzed in the model is more suitable (Mannering and Bhat, 2014;Mannering et al, 2016). Many authors have utilized and compared Bayesian models in univariate and multivariate settings, proving that multivariate models are superior to univariate models because they capture the within-area correlation among different crash severities (see for example, Aguero-Valverde, 2013; Barua et al, 2014;Boulieri et al, 2016;Wu and Donnell, 2016;Liu and Sharma, 2018;Y. Wang and Kockelman, 2013).…”

Section: Statistical Methods Overviewmentioning

confidence: 99%

“…From a methodological standpoint, a wide range of macro-level methods has been utilized in the literature. These methods can be classified into (a) classical models, including NB models ( -Aty and Radwan, 1998;Chin and Quddus, 2003;Hadayeghi et al, 2003;Noland and Oh, 2004;Noland and Quddus, 2004;Poch and Mannering, 1996), loglinear models (Abdel-Atyet al, 1998;Lee et al, 2006;Washington, 2006), (b) spatial autoregressive models (LaScala, et al, 2000;Levine et al, 1995;Quddus, 2008;Rhee et al, 2016), and (c) full Bayesian or hierarchical models (Aguero-Valverde, 2013;Aguero-Valverde andJovanis, 2010, 2006a;Boulieri et al, 2016;Cheng et al, 2018;Lee et al 2015;Liu and Sharma, 2018;Miaou, 1994;Siddiqui, et al, 2012;Truong et al, 2016). In the following sections, literature that has used each class of methods described above is summarized.…”

Section: Macroscopic Traffic Safety Researchmentioning

confidence: 99%

See 1 more Smart Citation

Bayesian Spatio-Temporal Analysis of Road Traffic Crashes

Azimian¹

View full text Add to dashboard Cite

BAYESIAN SPATIO-TEMPORAL ANALYSIS OF ROAD TRAFFIC CRASHES Amin Azimian Road traffic crashes are one of the major causes of death and serious injury in the US, leading to economic losses and human suffering. In recent years, several research efforts have been made to screen areas and locate hotspots/zones and to identify the factors contributing to traffic crashes of different severity levels. Such research typically aggregates crash locations into spatial units at the macro level, such as counties, or at the micro level, such as road segments or intersections. Many crash estimation methods have been proposed in the literature. These methods range from classical approaches, such as the linear, Poisson, negative binomial, and logistic regression methods, to more state-of-the-art approaches, such as the empirical Bayesian (EB), spatial autoregressive, and full Bayesian methods. A considerable drawback of classical methods is that they cannot account for the regression-to-the-mean bias and potential unobserved heterogeneity, which can result in unstable and biased parameter estimates. Additionally, EB and spatial autoregressive methods are unable to address multilevel data and group-level random effects. By contrast, the full Bayesian framework is more flexible and can be easily extended to include random effect terms that can act as proxies for unobserved or missing covariates that have a spatial or temporal structure. Although the full Bayesian framework appears to be a promising methodology for dealing with crash data, the few previous studies in the area of road traffic safety that used this framework had some major limitations. Some studies performed their analyses in DEDICATION This dissertation is lovingly dedicated to my family. Your support and encouragement mean so much to me, and I am forever grateful. Thank you! v ACKNOWLEDGMENTS First, I would like to thank my advisor, Dr. V. Dimitra Pyrialakou. I am deeply honored to be her first Ph.D. student. I appreciate all her contributions in terms of her time, ideas, and efforts to make my Ph.D. experience truly productive and stimulating. The joy and enthusiasm she has for research were contagious and motivational for me, especially during the tough times in my Ph.D. pursuit. I also thank my committee members, namely, Drs. David Martinelli, Kakan Dey, Sijin Wen, Ashish Nimbarte, and Steven Lavrenz, for all their advice and support. Finally, I thank my family for their love, care, and support. Specifically, I express my immense gratitude to my lovely wife, Hanieh, for her patience and unparalleled support. My family is my biggest support system, materially and morally. vi

show abstract

Section: Statistical Methods Overviewmentioning

confidence: 99%

Section: Macroscopic Traffic Safety Researchmentioning

confidence: 99%

Bayesian Spatio-Temporal Analysis of Road Traffic Crashes

Azimian¹

View full text Add to dashboard Cite

show abstract

“…Ghosh et al () provide a comprehensive discussion of the use of generalized linear modeling in Bayesian hierarchical small area estimation. Notable examples of a space‐time Bayesian hierarchical approach for modeling count data include modeling disease risk (Knorr‐Held ), election polls (Shor et al ), burglary (Li et al ), cycle accidents (DiMaggio ), and road traffic accidents (Boulieri et al ).…”

Section: Spatiotemporal Estimation and Forecasting For Public Opinionmentioning

confidence: 99%

A Spatiotemporal Bayesian Hierarchical Approach to Investigating Patterns of Confidence in the Police at the Neighborhood Level

Williams

Haworth

Blangiardo

et al. 2018

Geographical Analysis

Self Cite

View full text Add to dashboard Cite

Public confidence in the police is crucial to effective policing. Improving understanding of public confidence at the local level will better enable the police to conduct proactive confidence interventions to meet the concerns of local communities. Conventional approaches do not consider that public confidence varies across geographic space as well as in time. Neighborhood level approaches to modeling public confidence in the police are hampered by the small number problem and the resulting instability in the estimates and uncertainty in the results. This research illustrates a spatiotemporal Bayesian approach for estimating and forecasting public confidence at the neighborhood level and we use it to examine trends in public confidence in the police in London, UK, for Q2 2006 to Q3 2013. Our approach overcomes the limitations of the small number problem and specifically, we investigate the effect of the spatiotemporal representation structure chosen on the estimates of public confidence produced. We then investigate the use of the model for forecasting by producing one-step ahead forecasts of the final third of the time series. The results are compared with the forecasts from traditional time-series forecasting methods like na€ ıve, exponential smoothing, ARIMA, STARIMA, and others. A model with spatially structured and unstructured random effects as well as a normally distributed spatiotemporal interaction term was the most parsimonious and produced the most realistic estimates. It also provided the best forecasts at the London-wide, Borough, and neighborhood level.

show abstract

“…Recent examples include Boulieri et al . () and Eksler (). The former is a report on an aggregate areal analysis of road injury rates in the United Kingdom, and the latter, an areal analysis of road injury rates in Europe.…”

Section: Introductionmentioning

confidence: 99%

“…Nevertheless, even when spatial information is taken into account, many published analyses of road collisions regard road accidents as if they occur in unconstrained two-dimensional space. Recent examples include Boulieri et al (2016) and Eksler (2008). The former is a report on an aggregate areal analysis of road injury rates in the United Kingdom, and the latter, an areal analysis of road injury rates in Europe.…”

Section: Introductionmentioning

confidence: 99%

Using hidden Markov models to model spatial dependence in a network

Kadhem

Hewson

Kaimi

2018

Aus NZ J of Statistics

View full text Add to dashboard Cite

Summary This study considers spatial dependence in the number of injury crashes reported on a road network. The aggregated crash counts are considered realisations of a Poisson random variable; thus, we model both over‐dispersion and serial correlation using the Poisson hidden Markov model (PHMM). PHMMs have typically been used for modelling temporal dependence, but they have rarely been used to model spatial dependence. Our interest, however, is specifically in relation to an underlying point process which is constrained to occur on a network. We illustrate the use of the PHMM with police‐reported data on injury road collisions on selected motorways in the United Kingdom over a 5‐year period (2010–2014). We use officially recorded estimates of traffic volume as an exposure variable. The aim is to identify highway segments which might have distinctly high crash rates. To do this, we first select an optimal model in terms of the number of latent states. As we use a Bayesian approach, we can assign a posterior classification probability to each segment in terms of membership of the various underlying risk states. Model fitting is conducted using the Markov chain Monte Carlo approach; we develop a new modified version of the Akaike Information Criterion and Bayesian Information Criterion, approximated from a Bayesian framework, to select the best model in terms of number of states.

show abstract

A Space–Time Multivariate Bayesian Model to Analyse Road Traffic Accidents by Severity

Cited by 42 publications

References 69 publications

Bayesian Spatio-Temporal Analysis of Road Traffic Crashes

Bayesian Spatio-Temporal Analysis of Road Traffic Crashes

A Spatiotemporal Bayesian Hierarchical Approach to Investigating Patterns of Confidence in the Police at the Neighborhood Level

Using hidden Markov models to model spatial dependence in a network

Contact Info

Product

Resources

About