Quantum NLP for Predicting Injury Leaves in Occupational Accident Reports

Vol 56, 2024 - 309903
Resumo Estendido
Favoritar este trabalho
Como citar esse trabalho?
Resumo

Work accidents can threaten workers’ health and impose significant costs on organizations, such as the need for work restructuring and the direct or indirect costs associated with employee absence. Accident investigation reports offer valuable information that can help companies propose preventive measures and identify the causes and consequences of injury events. However, this information is often complex, redundant, or incomplete, making thorough human review challenging due to the large volume of reports. In this study, we applied Quantum Natural Language Processing (QNLP) techniques to predict injury leave based on accident reports from a hydroelectric power company. Our cross-validation results showed a median accuracy of 73.33%. We also analyzed reports with high and low classification accuracy to explore potential reasons for these outcomes. This study demonstrates that accident investigation reports can provide essential insights to enhance workplace safety.

Compartilhe suas ideias ou dúvidas com os autores!

Sabia que o maior estímulo no desenvolvimento científico e cultural é a curiosidade? Deixe seus questionamentos ou sugestões para o autor!

Faça login para interagir

Tem uma dúvida ou sugestão? Compartilhe seu feedback com os autores!

Instituições
  • 1 Universidade Federal de Pernambuco
  • 2 UFPE/CEERMA
  • 3 CEERMA/NT-CAA/UFPE
  • 4 CEERMA/DEP/UFPE
Eixo Temático
  • 26. SE-QPO - Métodos Quânticos e suas Aplicações em Pesquisa Operacional
Palavras-chave
Quantum Computing
Natural Language Processing
Occupational Safety
Accident Report