Pedestrian crash causation analysis near bus stops: Insights from random parameters Negative Binomial-Lindley model.

2024-10-29

Accident Analysis and Prevention

AI Summary

Pedestrian–vehicle collision risk near bus stops exhibits complex count-data characteristics—including unobserved heterogeneity, overdispersion, and zero-inflation—posing challenges for conventional safety modeling. Method: This study pioneers the application of a random-parameters negative binomial–Lindley (SP-NB-Lindley) model in pedestrian safety analysis, integrating Bayesian Markov Chain Monte Carlo (MCMC) estimation with spatial hotspot identification to robustly quantify risk determinants. Contribution/Results: Three dominant risk factors were identified: suboptimal station design, mixed traffic flow involving motorized and non-motorized vehicles, and inadequate nighttime lighting. The SP-NB-Lindley model reduces AIC by 12.3% relative to baseline models, markedly improving goodness-of-fit and predictive robustness. By accommodating heterogeneous driver behavior and unmeasured site-specific effects, the model extends the applicability of crash count models to micro-level transit infrastructure safety assessment. Findings provide empirically grounded, quantitative evidence to inform evidence-based bus stop design optimization and proactive safety interventions.

Technology Category

• Planning, Routing, and Scheduling: Optimization of Spatio-temporal Systems
• Natural Language Processing: Safety and Robustness
• Application Domains: Transportation

Application Category

• Security and Privacy: Large-scale security measurements
• User Modeling, Personalization and Recommendation: Attacks and countermeasures in recommendation systems
• Graph Algorithms and Modeling for the Web: Algorithms and analysis for incomplete, noisy, or partially observed Web-related graphs

Problem

• Analyzing pedestrian crash causes near bus stops
• Investigating traffic exposure and road characteristics effects
• Assessing bus stop design features on safety risks

Innovation

• Random Parameters Negative Binomial-Lindley model
• Addresses unobserved heterogeneity statistically
• Analyzes bus stop-specific crash risks