Jove
Visualize
Contact Us
JoVE
x logofacebook logolinkedin logoyoutube logo
ABOUT JoVE
OverviewLeadershipBlogJoVE Help Center
AUTHORS
Publishing ProcessEditorial BoardScope & PoliciesPeer ReviewFAQSubmit
LIBRARIANS
TestimonialsSubscriptionsAccessResourcesLibrary Advisory BoardFAQ
RESEARCH
JoVE JournalMethods CollectionsJoVE Encyclopedia of ExperimentsArchive
EDUCATION
JoVE CoreJoVE BusinessJoVE Science EducationJoVE Lab ManualFaculty Resource CenterFaculty Site
Terms & Conditions of Use
Privacy Policy
Policies

Related Experiment Videos

A 'Cluster-then-Estimate' Natural Language Processing (NLP) Approach for Classifying Maritime Incident Severity Based

Tianyi Chen1, Maohan Liang2, Wei Siong Lee1

  • 1Department of Civil and Environmental Engineering, National University of Singapore, Singapore 117576.

Accident; Analysis and Prevention
|January 23, 2026
PubMed
Summary

Related Concept Videos

You might also read

Related Articles

Articles linked to this work by shared authors, journal, and citation graph.

Sort by
Same author

Development and validation of a novel multimodal deep neural network model based on CBC digit parameters and scattergrams for rapid hematolymphoid malignancy classification: a multicenter cohort study.

BMC medicine·2026
Same author

A bioactive lysate of Corynebacterium glutamicum orchestrates intestinal barrier repair mainly through L-arginine-mediated FOXO pathway.

Food research international (Ottawa, Ont.)·2026
Same author

Design of M13 phage systems dual-displaying two functional proteins for developing a rapid biosensing probe.

Journal of nanobiotechnology·2026
Same author

Gut-Derived Lipid Mediators Orchestrate Ovarian Metabolic Homeostasis and Clutch Persistence in Aging Laying Hens via the PLA2G6-ALOX15B-AGPAT3 Axis.

Biomolecules·2026
Same author

Research advances in the identification of gut keystone bacteria and formulation of suitable dietary intervention strategies for restoring gut homeostasis.

NPJ biofilms and microbiomes·2026
Same author

A two-stage approach for the identification and classification of osteoporotic vertebral compression fractures.

Medical physics·2026
This summary is machine-generated.

This study introduces a novel NLP approach using Latent Dirichlet Allocation (LDA) and Bidirectional Encoder Representation from Transformers (BERT) to automatically estimate maritime incident severity from text. The method significantly improves accuracy in assessing vessel risk and managing incident data.

Area of Science:

  • Maritime Safety
  • Natural Language Processing
  • Data Science

Background:

  • Manual estimation of maritime incident severity from textual descriptions is inefficient for large datasets.
  • Current methods lack the speed and accuracy needed for effective risk assessment and historical data management.

Purpose of the Study:

  • To develop and validate an automated approach for estimating maritime incident severity using Natural Language Processing (NLP).
  • To enhance the efficiency and accuracy of incident severity assessment in the maritime industry.

Main Methods:

  • A 'cluster-then-estimate' strategy was employed, utilizing Latent Dirichlet Allocation (LDA) for text preprocessing and clustering.
  • Bidirectional Encoder Representation from Transformers (BERT) models were fine-tuned within each cluster for severity estimation.
Keywords:
Maritime incident assessmentNatural Language Processing (NLP)incident severity estimationtext classificationtext clusteringtextual incident description

Related Experiment Videos

  • A dataset of 22,458 maritime incidents, categorized by severity, was used for training and validation.

    • Main Results:

    • The proposed 'cluster-then-estimate' approach demonstrated superior performance compared to state-of-the-art baseline models.
    • The method accurately estimates incident severity, outperforming existing techniques.
    • Fine-tuning BERT within LDA-generated clusters significantly enhanced estimation capabilities.

    Conclusions:

    • The NLP-based 'cluster-then-estimate' approach offers a practical and valuable solution for automated incident severity estimation.
    • This methodology provides significant benefits for improving maritime incident assessment and decision-making processes.
    • The study represents a pioneering application of NLP for maritime incident severity analysis based on textual data.