Which approach better samples extreme traffic conflicts? Conventional- vs. machine learning-based sampling methods.
Clinical Snapshot
PICO Framework
| P — Population | Vehicle-to-pedestrian conflicts at urban signalized intersections |
| I — Intervention | Machine learning-based sampling methods (autoencoder neural network and isolation forest) |
| C — Comparator | Conventional sampling techniques with predefined thresholds |
| O — Outcomes | Alignment with conceptual severity level framework and compatibility with extreme value distribution |
Bottom Line
This methodological study demonstrates that machine learning approaches, particularly isolation forest algorithms, outperform conventional threshold-based methods for sampling extreme traffic conflicts in crash risk assessment. The research addresses a critical gap in extreme value theory application for proactive road safety analysis. While the findings show promise for improving conflict-based crash prediction models, the study lacks quantitative effect measures and statistical testing. The methodology could enhance Australian road safety programs, but validation with local traffic data would be needed before implementation. Transportation engineers should consider these ML approaches when developing conflict-based safety assessment tools, though conventional methods remain viable pending further validation studies.
Key Findings
P Value: Not reported
Effect Size: Qualitative superiority demonstrated but not quantified
Primary Outcome: Machine learning methods better aligned with conceptual severity levels than conventional sampling
Nnt Or Sensitivity: Isolation forest showed best preservation of empirical tail distribution characteristics
Confidence Interval: Not reported
Clinical Application
Requires access to traffic conflict databases and ML computational resources Relevant to Australian road safety initiatives and VicRoads/Transport for NSW conflict analysis programs, though validation needed for local traffic conditions Traffic safety researchers and transportation engineers working with conflict-based crash risk assessment
Abstract
Extreme value theory has been receiving much attention of late for proactively estimating crash risk through a two-step procedure that first samples extreme traffic conflicts and then estimates crash risk based on those sampled extremes. Although the existing body of research has encapsulated sampling methods within a predominant conventional technique, there is no universally accepted practice on how to efficiently select threshold values, nor on how to evaluate the sampled extreme conflicts alignment with the conceptual crash severity level framework. This research aims to address these issues by employing machine learning-based sampling methods, which do not require predefined thresholds, and by comparing the sampled extremes with the conceptual severity levels, to assess their alignment. After a review of recent developments in machine learning techniques in transportation and other engineering fields, two promising machine learning sampling models, autoencoder neural network and isolation forest, were investigated using a database of vehicle-to-pedestrian conflicts at urban signalized intersections. Sampled extreme conflicts using the machine learning and conventional sampling techniques-as a baseline -were assessed and compared using two criteria: their visual alignment with the conceptual severity level framework, and their compatibility with the extreme value distribution. The results demonstrate that the extreme conflicts selected based on the machine learning methods better mirror the conceptual severity levels than the conventional sampling technique. Moreover, extremes classified by the isolation forest more closely preserve the characteristics of the empirical tail distributions, demonstrating a better contextual representation for modeling with the extreme value distribution compared to the autoencoder neural network and conventional sampling methods.
References
- 1.Hasanpour, M., Chen, Z., D'Agostino, C., Persaud, B., & Milligan, C. (2026). Which approach better samples extreme traffic conflicts? Conventional- vs. machine learning-based sampling methods. Accident Analysis and Prevention, 108423. https://doi.org/10.1016/j.aap.2026.108423
This content is for educational purposes for healthcare professionals only and does not constitute clinical advice. Clinical decisions should be based on individual patient assessment, current guidelines, and appropriate specialist consultation. Editorial Standards · Privacy Policy · Terms of Service