Review of: "Crash severity analysis of vulnerable road users using machine learning"
DOI: 10.32388/uuik52
archive: archived pipeline: cataloged verified
Get this paper ↗ (DOI — opens at the source; we link to it, we don't host it)
Summary
This document is a peer review of a manuscript titled "Crash severity analysis of vulnerable road users using machine learning," submitted to PLOS ONE. The reviewer, Apostolos Ziakopoulos from the National Technical University of Athens, recommends acceptance with major revisions. The reviewed paper investigates the injury severity of vulnerable road users (VRUs) using various machine learning algorithms for binary classification. The reviewer commends the study for its appropriate methodologies and clear writing, noting that while the results are useful, they are not extremely novel. The review identifies several areas requiring improvement in the manuscript’s structure and content. The reviewer suggests condensing the abstract and expanding the literature review to include specific details on algorithmic variations and performance indicators from prior studies. Four specific recent studies are recommended for inclusion to strengthen the contextual background, covering topics such as spatio-temporal crash patterns, binary logistic regression modeling, weather and traffic data impacts, and in-depth European studies on two-wheeler crashes. Additionally, the reviewer advises moving introductory text regarding classification approaches from the methodology section to the literature review to improve flow. Substantive methodological concerns are raised regarding data clarity and variable coding. The reviewer points out an inconsistency in the description of the Queensland (QLD) crash database, specifically regarding the recording of injury versus non-injury crashes between 2010 and 2019, urging clarification on whether the QLD and TMR databases differ. Furthermore, the reviewer critiques the binary coding of vehicle and driver conditions in Table 1, arguing that categories such as helmet use and inattention are not mutually exclusive in reality. The reviewer suggests exploring alternative coding methods and adding 100% totals to the table for clarity. Technical details regarding hyperparameter tuning are also requested, including the number of combinations tested for algorithms other than KNN and the runtime required for the grid search. The reviewer highlights issues with result interpretation and presentation. Figure 4’s reference to Gini importance requires explicit labeling, and the counter-intuitive finding in Figure 5—that younger ages correlate with lower severe crash probabilities—needs elaboration to reconcile with existing literature. Finally, the reviewer notes that the models’ medium performance and low specificity are not adequately addressed in the limitations section. The authors are urged to rewrite this section to discuss how model utility can be improved. Minor English language revisions are also recommended to enhance readability.
Provenance
The full processing record for this entry. Every stage of this paper's journey through the pipeline is logged — what ran, with which tool and model, how many attempts it took, and when it last completed.
| Stage | Outcome | Tool | Model | Prompt | Attempts | Completed |
|---|---|---|---|---|---|---|
| discover | success | Crossref | — | — | 1 | 2026-06-18 |
| archive | success | canonical_url | — | — | 1 | 2026-06-25 |
| extract | success | pdftotext | — | — | 2 | 2026-06-26 |
| clean | success | clean | — | — | 1 | 2026-06-26 |
| chunk | success | chunk | — | — | 1 | 2026-06-26 |
| embed | success | embed | Qwen/Qwen3-Embedding-8B | — | 1 | 2026-06-26 |
| enrich | success | semantic_scholar | — | — | 4 | 2026-06-26 |
| promote | success | — | — | — | 1 | 2026-06-18 |
| summarize | success | llm | qwen3.6-27b-prismaquant | summ-v5 | 1 | 2026-06-26 |
| tag | success | vector_similarity | — | — | 6 | 2026-06-26 |
| verify | success | — | — | — | 1 | 2026-06-26 |
Summary generated by qwen3.6-27b-prismaquant on 2026-06-26; verification: verified.
Topics
Ranked by relevance to this paper. Hover a topic for its definition.
- vru crash typology
- motorcycle crash typology
- telematics crash prediction
- demographic disparities
- motorcyclist skill
Information type
What kind of knowledge this paper contributes, grouped by family — independent of topic (what it is about) and method (how it was studied).
- Empirical Findings: crash risk outcomes