Sequence Labeling-based Multiple Causal Relations Extraction using Pre-trained Language Model for Maritime Accident Prevention 


Vol. 38,  No. 5, pp. 51-56, Oct.  2023
10.14346/JKOSOS.2023.38.5.51


PDF
  Abstract

In the inspection of workplace hazards/risk factors by specialized institutions dedicated to safety management, inspection reports vary based on the inspectors, who lack the authority to enforce improvement of workplace hazards/risk factors. Thus, improvement and accident rates remain steady without decreasing. This study performed a regression analysis on the relationship between improvement and accident rates of categorized inspection items by classifying hazards/risk factors from inspection reports submitted by a specialized safety management institution in Chungbuk after inspecting 10 food and beverage manufacturers over the past three years. The hazards/risk factors were classified into five categories: mechanical, electrical, chemical, human, and environmental. The regression analysis revealed that the improvement rate of hazards/risk factors inspected by the specialized safety management institution influenced accident rates. To enhance improvement rates based on these findings, this study prioritized the correction of the five most frequently cited inspection items with the lowest improvement rates in each area. Based on these inspection items, this study suggested a checklist for use in workplace safety inspections of food manufacturers. This proposed checklist is expected to reduce accident rates in food manufacturing facilities. Currently, guidance and inspection of workplaces are mainly focused on accident rates rather than correcting hazards/risks. Thus, accident rates remain unchanged as workplace risks are inadequately improved according to the unique characteristics of each workplace. When conducting workplace guidance and inspection, policy measures and inspection methods are warranted to increase the improvement rate of hazards/risks.

  Statistics
Cumulative Counts from November, 2022
Multiple requests among the same browser session are counted as one view. If you mouse over a chart, the values of data points will be shown.


  Cite this article

[IEEE Style]

문기영, 김도현, 양태훈, 이상덕, "Sequence Labeling-based Multiple Causal Relations Extraction using Pre-trained Language Model for Maritime Accident Prevention," Journal of the Korean Society of Safety, vol. 38, no. 5, pp. 51-56, 2023. DOI: 10.14346/JKOSOS.2023.38.5.51.

[ACM Style]

문기영, 김도현, 양태훈, and 이상덕. 2023. Sequence Labeling-based Multiple Causal Relations Extraction using Pre-trained Language Model for Maritime Accident Prevention. Journal of the Korean Society of Safety, 38, 5, (2023), 51-56. DOI: 10.14346/JKOSOS.2023.38.5.51.