Aug 2026· Accident Analysis and Prevention· Vol 236, pp.
108705
· 1 citation· 75 references
Medicine
Abstract
Road traffic crashes remain a major global safety concern, and segment-level crash analysis plays a critical role in identifying high-risk locations. However, such analyses are highly sensitive to spatial unit definitions, giving rise to the Modifiable Areal Unit Problem (MAUP). Existing studies predominantly rely on fixed or single-scale segmentation, which limits their ability to capture scale-dependent effects and may bias both model estimation and hotspot identification. To address this issue, this study proposes a hybrid multi-scale road segmentation framework that integrates roadway homogeneity with crash distribution characteristics. The framework employs a Poisson likelihood ratio test and leave-one-out cross-validation (LOOCV) to generate adaptive segmentation schemes, and evaluates MAUP effects through crash distribution analysis, negative binomial modeling, and external validation. The results show that segmentation choice materially affects statistical representation, model estimation, and predictive performance. Both the scale effect and the zoning effect of MAUP are found to influence crash modeling and hotspot-related inference, although their impacts are not identical. External validation reveals substantial performance differences among segmentation schemes, with Seg-6 showing the strongest predictive performance within the original parameter set; the sensitivity analysis further indicates that this result is locally robust within the evaluated parameter neighborhood. Major crash concentration patterns remain broadly stable across segmentation schemes, whereas minor local variations are more segmentation-sensitive. These findings show that segmentation should be treated as an explicit analytical design issue rather than a neutral preprocessing step, and provide a systematic basis for evaluating MAUP effects in segment-level traffic safety analysis.
Introduction: Crash counts on road segments and intersections exhibit differ- ent exposure and connectivity patterns that conventional analyses may obscure. Methodology: A Bayesian negative binomial node edge model was fitted to 8,169 road segments and 8,398 intersections in six central districts of Bogot\'a. Separate predictors represented road hierarchy, pavement, speed, signalization, intersection configuration, and land-use treatment. Model performance was examined through pre- dictive summaries and spatial diagnostics, while computational details are reported in the appendix. Results: Intersections with at least four incident segments and higher maximum incident speeds had higher expected crash counts. Land use treatment and signalized access intensity also showed posterior associations. Road hierarchy and signalization were the clearest segment-level factors; however, this component had weak raw scale predictive performance and numerical uncertainty for some pavement categories. Conclusion: Treating intersections and segments as distinct network elements provides an interpretable baseline for urban crash analysis, but the segment results require cautious interpretation and motivate spatially structured extensions.
Danna Lesley Cruz Reyes, Cristian Harvey Ardila Bolívar· 0 citations
Road traffic accidents on national highways pose a significant public health and economic challenge in Bangladesh, necessitating systematic safety assessment. This study analyzes accident trends, contributing factors, and spatial patterns by identifying accident-prone locations (blackspots) along the Kushtia–Jhenaidah National Highway (N704). Accident data for the period 2017–2021 were obtained from nearby police stations. In addition, a cluster random sampling approach was used to conduct a questionnaire survey involving 100 participants, including drivers and general road users, to capture behavioural insights related to accident occurrence. The study integrates descriptive statistical methods, such as trend analysis and frequency distribution, with spatial techniques including severity index evaluation, Kernel Density Estimation (KDE), and hotspot analysis.The findings indicate a decline in overall accident frequency from 2018 to 2021, while fatality rates increased in 2021. Heavy vehicles, particularly trucks, were identified as major contributors to accidents, and head-on collisions emerged as the most common crash type. Key risk factors include driver inexperience, mobile phone usage while driving, overspeeding, inadequate training, and nighttime driving conditions. The analysis further reveals that individuals aged 20–40 are the most affected group, with higher fatality rates among males and higher injury rates among females.A total of 35 accident-prone locations were identified, with several segments classified as blackspots based on accident frequency, injury severity, and fatality occurrence. The study recommends targeted interventions such as driver training, infrastructure improvement, enhanced enforcement, and coordinated policy actions to improve highway safety and reduce accident risks.
N. O, Bhagyalakshmi, Surendrababu M S et al.· International Journal of Res...· 0 citations
Bulk carriers play a critical role in global dry bulk transportation, and their safe operation is closely related to commodity supply chains, port continuity, and maritime governance. However, bulk carrier accidents are unevenly distributed across maritime space, and existing maritime blackspot studies are often limited by single-scale density estimation, unconstrained planar smoothing, and insufficient consideration of temporal persistence. These limitations make it difficult to distinguish robust accident-prone waters from scale-sensitive or temporally unstable hotspots. To address this problem, this study proposes a constrained multiscale consensus framework for identifying and interpreting global bulk carrier accident blackspots. The framework first screens and standardizes global maritime accident records to extract valid bulk carrier accident samples. It then constructs an ocean-constrained equal-area analysis grid and estimates severity-weighted accident intensity under multiple Gaussian smoothing bandwidths. Scale-specific hotspots are further extracted through threshold-based segmentation and minimum-area filtering, and a consensus persistence rule is developed to classify core, secondary, and transition blackspots. Finally, threshold sensitivity analysis, bootstrap resampling, time-window comparison, lifecycle classification, accident-type stratification, and severity-weighted versus frequency-only comparison are conducted to evaluate the robustness and interpretability of the identified blackspots. Based on 38,139 raw accident records, the empirical analysis retained 1441 cleaned bulk carrier accidents from 2015 to 2023 and identified 87 core consensus blackspots, covering approximately 8.06 million km2 and containing 863 accidents. These blackspots are mainly concentrated in major coastal shipping regions, and the proposed framework provides a reproducible and geographically constrained basis for global maritime blackspot identification.
Zhanzhu Li, Xiaohua Cao, Jin Chen et al.· Journal of Marine Science an...· 0 citations
OBJECTIVE
Highway work zones introduce spatial and operational disruptions that elevate crash risk. Although prior research has examined work zone crashes at an aggregated level, less attention has been given to how crash types vary across functional work zone areas (Advance Warning, Transition, Activity, and Termination), where operating conditions and risk differ by segment. This study examines crash-type differentiation conditional on crash occurrence.
METHODS
Using a multi-state work zone crash dataset of 20,617 records, segment-specific multinomial logistic regression classifiers were developed to classify five dominant crash types (rear-end, sideswipe, run-off-road, head-on/front, and rollover/overturn) and to identify the environmental and temporal factors associated with crash-type differentiation within each segment.
RESULTS
Segment-specific models showed moderate and stable classification performance, with overall accuracies ranging from 0.74 to 0.77 across the four work-zone segments. Fold-level cross-validation summaries and state-specific testing were conducted to further assess model robustness. Crash-type-specific risk-factor assessment revealed that lighting and surface conditions consistently influenced crash patterns across segments, while weather, season, and time of day exerted segment- and crash-type-dependent effects.
CONCLUSIONS
Segment-specific crash-type classification provides a context-aware basis for understanding heterogeneous crash mechanisms within work zones. The identified segment- and crash-type-dependent risk profiles support the design of targeted countermeasures (e.g., visibility enhancement, surface treatments, and time-responsive traffic control) to improve work zone safety management.
Traditional identification of accident-prone road sections relies on fixed-unit statistics, which has disadvantages that include missing spatial continuity of risk, boundary distortion, and information “averaging,” and struggles to support differentiated safety management. To address this issue, this study proposes a road accident risk field reconstruction model integrating dynamic spatial attenuation and probability-severity dual-dimensional coupling. First, based on the information diffusion principle and Gaussian kernel density estimation (KDE), a framework for converting discrete accident points into a continuous risk probability field is constructed, and a dynamic standard deviation function is proposed. This function enables the Gaussian kernel standard deviation to dynamically adapt to road design speed and cross-sectional type (with/without median divider), clarifying its physical correlation with drivers’ sight distance requirements. Second, by integrating casualties and direct economic losses, a continuous accident severity field is established through an accident equivalent loss function; the two fields are discretized using the quantile method, and combined with a risk matrix to generate comprehensive risk ratings (Levels I–V) via coupling. Verification using accident data from a Class II mountain highway in Guangxi shows the following: (1) the boundaries of the risk field generated by the model fit the road alignment well, with a probability field gradient smoothness of 0.0068, which can mitigate the boundary effect and step effect of the fixed-unit method; (2) compared with KDE with fixed bandwidth (0.35 km) and fixed-length segmentation method (0.5 km), the area under the curve (AUC) of the model’s probability field reaches 0.9641 (increased by 12.79% and 15.31%, respectively), the AUC of the severity field is 0.8669 (increased by 15.24% and 5.88%, respectively), and the Pearson correlation coefficient between the probability field and accident data is 0.4772 (increased by 11.01% and 3.77%, respectively), indicating that the identification accuracy and data fit have been improved; and (3) dual-dimensional coupling can distinguish risk patterns of high-frequency low-loss, low-frequency high-loss, and high-frequency high-loss, providing a quantitative basis for differentiated management. This model promotes the evolution of road risk assessment from discrete statistics to continuous field analysis, and can offer technical support for the optimal allocation of safety resources.
Jianfeng Liu, Fuyuan Luo, Jianqiu Chen et al.· Journal of Transportation En...· 0 citations
OBJECTIVE
This study aimed to address key data limitations in autonomous vehicle (AV) crash-severity analysis, including small samples and sample imbalance, and to identify interpretable risk factors associated with injury outcomes in AV crashes.
METHODS
A verified dataset of 2,946 AV crash events from 2015 to 2024 was constructed, with crash records spatially matched to variables related to road geometry, traffic control, roadway attributes, and the built environment. A hybrid analytical framework was developed by combining sample-balancing methods, feature-selection techniques, and random-parameter logit modeling. Three balancing strategies, ROSE, SMOTE, and ROSE+SMOTE, were compared with three feature-selection methods, mutual information (MI), random forest (RF), and XGBoost, under both balancing-first and feature-selection-first workflows. Nineteen model specifications were evaluated using stratified five-fold cross-validation. Further causal analysis of AV crash severity was conducted based on the parameter estimates and marginal effects of the random-parameter logit model.
RESULTS
All hybrid specifications improved performance relative to the unprocessed baseline. The combined ROSE+SMOTE strategy produced stronger performance than single balancing methods. Model 18, which applied ROSE+SMOTE before XGBoost feature selection, achieved the highest Macro-Recall, indicating strong balanced sensitivity across severity classes. Model 19, which applied XGBoost feature selection before ROSE+SMOTE, achieved the best overall and safety-oriented performance, with the highest Accuracy, Macro-Precision, Injury-class Recall, Macro-F1, ROC-AUC, PR-AUC, the lowest FNR-injury, and the lowest AIC. Significant injury-risk factors included side-impact crashes, restaurant density, school and metro stop, commercial and mixed-use/public land use, lane-markings-only, nighttime, rush hour, crosswalks, T/Y-intersections, expressways and arterials, intersection-related locations, and crash lanes ≤ 2. Greater road width was associated with a lower injury probability.
CONCLUSIONS
The proposed hybrid framework provides a stable and interpretable approach for analyzing small and imbalanced AV crash datasets, and offers evidence to support scenario-based ODD testing, targeted infrastructure improvement, curbside management, and context-sensitive AV safety governance.
Feng Tang, Ruien Wu, Ning Li et al.· Traffic Injury Prevention· 0 citations
We use cookies to run the site and, with your consent, for analytics and to show ads.
See our Cookie Policy.
