Structuring occupational accident narratives for cross-sector safety analysis: Transferability of accident-process role classification
Abstract
Introduction: Occupational accident narratives describe work situations, unfavourable conditions, accident events, and consequences, but differences in terminology and reporting practices hinder systematic analysis across sectors and organisations. This study examined whether a classification model developed in one occupational sector could identify the same accident-process information in unseen sectors and reporting environments. Method: French accident narratives were segmented into factual units and expert-annotated as work situation, explicitly reported unfavourable condition, accident event or deviation, or reported consequence. Models were developed on 42,244 factual units from 6,040 construction-sector narratives and evaluated without retraining on metallurgy, chemistry-plastics, and an independently collected company corpus. We compared a TF-IDF-based lexical model, frozen pretrained text representations, and task-adapted pretrained models using cross-entropy or supervised representation-learning objectives. Results: Average balanced accuracy was 75.0% for the TF-IDF-based model, 76.9% for frozen pretrained representations, and about 85.7% after task adaptation. Across repeated runs, the leading adapted approaches showed similar overall performance, with no method consistently outperforming the others. Performance was lower and more variable on the company corpus, where transfer also involved a different reporting environment and data source. Conclusions: Accident-process roles learned from construction narratives remained identifiable in other sectors and an independent organisational setting. Practical Applications: The framework can support safety professionals in the assisted coding and expert review of large accident-report collections. By structuring heterogeneous narratives into comparable accident-process elements, it can facilitate cross-sector analysis and help identify recurring patterns relevant to occupational accident prevention.
Related
- Essence and accident modalities meet Belnapian truth values
- Research on expressway traffic accident risk prediction method considering Spatio-Temporal effects
- Singular Equilibrium and Selection of Narratives
- The FlightFinder Occurrence Corpus: harmonising 128 aviation-safety source codes into one deduplicated cross-national accident dataset
- Time-Varying Data as Sheaves: an Invitation to Narratives
BibTeX
@misc{yapi2026structuring,
title = {{Structuring occupational accident narratives for cross-sector safety analysis: Transferability of accident-process role classification}},
author = {Aho Yapi and Pierre Latouche and Arnaud Guillin and Yan Bailly},
howpublished = {HAL},
year = {2026},
month = sep,
url = {https://hal.science/hal-05755309v2},
note = {CC BY 4.0},
}
RIS
TY - UNPB
TI - Structuring occupational accident narratives for cross-sector safety analysis: Transferability of accident-process role classification
AU - Aho Yapi
AU - Pierre Latouche
AU - Arnaud Guillin
AU - Yan Bailly
PY - 2026
DA - 2026/09/18
PB - HAL
UR - https://hal.science/hal-05755309v2
AB - Introduction: Occupational accident narratives describe work situations, unfavourable conditions, accident events, and consequences, but differences in terminology and reporting practices hinder systematic analysis across sectors and organisations. This study examined whether a classification model developed in one occupational sector could identify the same accident-process information in unseen sectors and reporting environments. Method: French accident narratives were segmented into factual units and expert-annotated as work situation, explicitly reported unfavourable condition, accident event or deviation, or reported consequence. Models were developed on 42,244 factual units from 6,040 construction-sector narratives and evaluated without retraining on metallurgy, chemistry-plastics, and an independently collected company corpus. We compared a TF-IDF-based lexical model, frozen pretrained text representations, and task-adapted pretrained models using cross-entropy or supervised representation-learning objectives. Results: Average balanced accuracy was 75.0% for the TF-IDF-based model, 76.9% for frozen pretrained representations, and about 85.7% after task adaptation. Across repeated runs, the leading adapted approaches showed similar overall performance, with no method consistently outperforming the others. Performance was lower and more variable on the company corpus, where transfer also involved a different reporting environment and data source. Conclusions: Accident-process roles learned from construction narratives remained identifiable in other sectors and an independent organisational setting. Practical Applications: The framework can support safety professionals in the assisted coding and expert review of large accident-report collections. By structuring heterogeneous narratives into comparable accident-process elements, it can facilitate cross-sector analysis and help identify recurring patterns relevant to occupational accident prevention.
ER -