Modeling Farm Equipment Vehicle crash injury severity using random parameters logit and multi-class support vector machine in a developing country.

Nasab, EJ; Sheikholeslami, A; Vassallo, JM; Moeinaddini, A · 2026 · PubMed Central

DOI: 10.1038/s41598-026-42267-8

archive: archived pipeline: cataloged verified

Get this paper ↗ (DOI — opens at the source; we link to it, we don't host it)

Summary

This study addresses the critical safety challenges posed by Farm Equipment Vehicles (FEVs) on interurban and rural roads in developing countries, specifically Iran. FEVs, such as tractors and harvesters, often lack modern safety features and operate under suboptimal conditions, leading to disproportionately high fatality rates compared to other rural traffic incidents. The research aims to model crash injury severity to identify key risk factors and evaluate predictive accuracy for Intelligent Transportation Systems. It fills a gap in existing literature by systematically modeling interaction effects and heterogeneity in FEV crash outcomes, which previous studies often overlooked. The researchers employed a dual-method approach using crash data from Iran. First, they applied a Random Parameters Ordered Logit (RPOL) model to capture both observed and unobserved heterogeneity in injury severity, treating severity as an ordinal variable with three levels: Property Damage Only, Injury, and Fatal. Second, they utilized a Multi-Class Support Vector Machine (MC-SVM) to benchmark predictive performance and identify non-linear relationships among crash factors. This combination allows for a comprehensive analysis of how driver behavior, vehicle characteristics, road geometry, and environmental conditions interact to influence outcomes. The RPOL results identified several significant factors influencing crash severity. Motorcycle involvement and nighttime crashes were found to significantly increase the risk of fatalities. Conversely, rear-end collisions and reversing maneuvers were associated with lower severity outcomes. The study also noted heterogeneity in collisions where FEVs were at fault, particularly on straight roads. In terms of predictive performance, the MC-SVM model demonstrated superior accuracy, achieving higher Area Under the Curve (AUC) scores compared to the RPOL model. These findings highlight that crash severity is rarely the result of isolated factors but emerges from complex interactions, such as the heightened danger of slow-moving vehicles on high-speed straight corridors. The study concludes that integrating statistical and machine learning approaches provides a robust framework for improving rural transportation safety. Policy implications include renewing aging FEV fleets, restricting older vehicles, and imposing stricter controls on nighttime operations. By identifying specific high-risk scenarios, such as nighttime motorcycle interactions and straight-road collisions, the findings support targeted interventions like improved visibility measures and infrastructure planning. This research contributes to the field by offering a more accurate predictive model for FEV crashes in developing nations, where outdated machinery and poor infrastructure exacerbate safety risks.

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.

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.

• demographic disparities
• incidence prevalence
• motorcycle crash typology
• vru crash typology
• comparative international
• induced exposure

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