Collective Risk Ranking of Highway Segments on the Basis of Severity-Weighted Crash Rates

Table 1

Highlights on Crash Risk Analysis (in chronological order).

Author	Domain	Methodology
Chiou and fu [9]	Joint models	Multinomial-generalized Poisson joint model with error components.
Sugiyanto et al. [10]	Collective risk	Upper Control Limit (UCL) for hotspots with equivalent accident number.
Ouni and Belloumi [11]	Risk factors	Multinomial logit model to identify factors affecting severity. Ripley’s K-function and kernel density estimation (KDE) for vulnerable road users (VRUs) clusters.
Xu et al. [12]	Risk factors crash severity	A sequential logistic regression model with random parameters for multiple collision angles.
Feng et al. [13]	Risk factors	Real-time experiment on the psychological and behavioral changes of drivers in longitudinal underpass segments.
Wen et al. [14]	Spatial correlation crash rate	A Poisson-based model with spatial conditional autoregressive priors (CAR) for crashes within adjacent road segments.
Zeng et al. [15]	Spatial correlation crash severity	Bayesian generalized ordered logit model with CAR prior spatial term, fixed parameters, and flexible thresholds.
Ulak et al. [16]	Spatial correlation crash rate	Getis-Ord Gi∗ and local Moran’s I distance-based spatial weights with crash prediction accuracy index (CPAI).
Wang et al. [17]	Risk factors	A questionnaire-based study on risk perception among cyclists, and the major influencing factors in the prediction of risky behaviors.