Statistical and machine-learning methods for clearance time prediction of road incidents: A methodology review

Tang, Jinjun; Zheng, Lanlan; Han, Chunyang; Yin, Weiqi; Zhang, Yue; Zou, Yajie; Huang, Helai

doi:10.1016/j.amar.2020.100123

Cited by 73 publications

(56 citation statements)

References 57 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The exclusive use of quantitative methods (specifically statistical modelling) to analyze risk factors suffers from several weaknesses when applied to injury studies. A reliance on large sample sizes/populations to achieve adequate statistical power biases studies of smaller locales towards the null hypothesis, almost certainly leading to important patterns and risk factors being rejected due to inadequate statistical significance [6][7][8][9]. A traditional hypothesis testing (or otherwise p-value or confidence-interval-focused) approach additionally relies on assumptions of underlying distributions to assume multiple samples from a consistent population (a particularly perilous problem when using geospatial models), which, well suited to inferential modelling of highly-controlled experimental conditions [10], falls short of accounting for the non-parametric nature of models.…”

Section: Mixed Methodsmentioning

confidence: 99%

“…Alcohol Outlets: Pedestrian injury has been associated with proximity to alcohol-serving establishments, possibly due to increased risk of intoxication by either drivers or their victims [7,35,36]. Many studies have demonstrated that pedestrians who are intoxicated sustain injuries at a greater rate and severity than those who are not [23,31,37,38].…”

Section: Risk Factorsmentioning

confidence: 99%

“…Kernel density estimation (KDE) was used to identify the chief hotspots using a bandwidth of 500 m (which approximates the average city block). KDE was selected based on its use in previous injury studies [7,48,49] and because it indicates location hotspots (based on hue) without revealing individual patient locations.…”

Section: Geographical Analysismentioning

confidence: 99%

“…We used a detailed literature review to identify chief correlates of PMVC to include in the survey instrument. The instrument was first used in a study conducted by the authors in Vancouver, BC [7]. Attributes included in the walking in-person environmental scans were: long block (Yes or No); bus stop (Y or N); curb parking (Y or N); cross-walk (Y or N); obstruction (Y or N); signage (high, medium, low, subjectively measured); lanes of traffic; L/R turn bans (Y or N); bars (Y or N); retail (Y or N); school (Y or N); traffic calming measures (Y or N); median (Y or N); exclusive turn lane (Y or N); and speed limit.…”

Section: Environmental Scansmentioning

confidence: 99%

See 3 more Smart Citations

Qualitative Field Observation of Pedestrian Injury Hotspots: A Mixed-Methods Approach for Developing Built- and Socioeconomic-Environmental Risk Signatures

Schuurman

Walker

Swanlund

et al. 2020

IJERPH

View full text Add to dashboard Cite

Road traffic injuries constitute a significant global health burden; the World Health Organization estimates that they result in 1.35 million deaths annually. While most pedestrian injury studies rely predominantly on statistical modelling, this paper argues for a mixed-methods approach combining spatial analysis, environmental scans, and local knowledge for assessing environmental risk factors. Using data from the Nova Scotia Trauma Registry, severe pedestrian injury cases and ten corresponding hotspots were mapped across the Halifax Regional Municipality. Using qualitative observation, quantitative environmental scans, and a socioeconomic deprivation index, we assessed hotspots over three years to identify key social- and built-environmental correlates. Injuries occurred in a range of settings; however, clear patterns were not observed based on land use, age, or socio-economic status (SES) alone. Three hotspots revealed an association between elevated pedestrian injury and a pattern of geographic, environmental, and socio-economic factors: low- to middle-SES housing separated from a roadside attraction by several lanes of traffic, and blind hills/bends. An additional generalized scenario was constructed representing common risk factors across all hotspots. This study is unique in that it moves beyond individual measures (e.g., statistical, environmental scans, or geographic information systems (GIS) mapping) to combine all three methods toward identifying environmental features associated with pedestrian motor vehicle crashes (PMVC).

show abstract

Section: Mixed Methodsmentioning

confidence: 99%

Section: Risk Factorsmentioning

confidence: 99%

Section: Geographical Analysismentioning

confidence: 99%

Section: Environmental Scansmentioning

confidence: 99%

See 2 more Smart Citations

Qualitative Field Observation of Pedestrian Injury Hotspots: A Mixed-Methods Approach for Developing Built- and Socioeconomic-Environmental Risk Signatures

Schuurman

Walker

Swanlund

et al. 2020

IJERPH

View full text Add to dashboard Cite

show abstract

“…Boosting is an approach for reducing bias. Outliers are highly tolerated, and the importance of explanatory variables can be evaluated in random forest models [28].…”

Section: Random Forest Algorithmmentioning

confidence: 99%

Quantifying the Effects of Visual Road Information on Drivers’ Speed Choices to Promote Self-Explaining Roads

Qin

Chen

Lin

2020

IJERPH

View full text Add to dashboard Cite

Roads should deliver appropriate information to drivers and thus induce safer driving behavior. This concept is also known as “self-explaining roads” (SERs). Previous studies have demonstrated that understanding how road characteristics affect drivers’ speed choices is the key to SERs. Thus, in order to reduce traffic casualties via engineering methods, this study aimed to establish a speed decision model based on visual road information and to propose an innovative method of SER design. It was assumed that driving speed is determined by road geometry and modified by the environment. Lane fitting and image semantic segmentation techniques were used to extract road features. Field experiments were conducted in Tibet, China, and 1375 typical road scenarios were picked out. By controlling variables, the driving speed stimulated by each piece of information was evaluated. Prediction models for geometry-determined speed and environment-modified speed were built using the random forest algorithm and convolutional neural network. Results showed that the curvature of the right boundary in “near scene” and “middle scene”, and the density of roadside greenery and residences play an important role in regulating driving speed. The findings of this research could provide qualitative and quantitative suggestions for the optimization of road design that would guide drivers to choose more reasonable driving speeds.

show abstract

An Improved Deep Neural Network-Based Predictive Model for Traffic Accident’s Severity Prediction

Simaiya

Lilhore

Pandey

et al. 2022

Ambient Communications and Computer Systems

View full text Add to dashboard Cite

Statistical and machine-learning methods for clearance time prediction of road incidents: A methodology review

Cited by 73 publications

References 57 publications

Qualitative Field Observation of Pedestrian Injury Hotspots: A Mixed-Methods Approach for Developing Built- and Socioeconomic-Environmental Risk Signatures

Qualitative Field Observation of Pedestrian Injury Hotspots: A Mixed-Methods Approach for Developing Built- and Socioeconomic-Environmental Risk Signatures

Quantifying the Effects of Visual Road Information on Drivers’ Speed Choices to Promote Self-Explaining Roads

An Improved Deep Neural Network-Based Predictive Model for Traffic Accident’s Severity Prediction

Contact Info

Product

Resources

About