lcampillos/medical-ner

Notebook for BERT medical named entity recognition

Medical Named Entity Recognition

Notebooks for medical named entity recognition with BERT and Flair, used in the article "A clinical trials corpus annotated with UMLS entities to enhance the access to Evidence-Based Medicine".

Update (2022): The annotated data and the BERT trained model is now available in the Huggingface hub.

Workspace requirements

• Python 3.6+
• Pytorch 1.3+
• Huggingface Transformers (tested with version 2.9)
• Flair (tested with version 0.6)
• Matplotlib (tested with version 3.2)
• Numpy (tested with version 1.17.4)
• Pandas (tested with version 0.20)
• Seaborn (tested with version 0.9)
• Seqeval (tested with version 0.12)
• A Graphical Processing Unit (GPU)

Acknowledgements

These code is inspired by the following resources:

https://github.com/Spain-AI/transformers (by Álvaro Barbero)

https://www.depends-on-the-definition.com/named-entity-recognition-with-bert/ (by Tobias Sterbak)

Check also the Flair tutorials at:

https://github.com/flairNLP/flair

How to cite

If you use this code, you can make reference to the article where the script was made available, as follows:

  A clinical trials corpus annotated with UMLS entities to enhance the access to Evidence-Based Medicine
  Leonardo Campillos-Llanos, Ana Valverde-Mateos, Adrián Capllonch-Carrión, Antonio Moreno-Sandoval
  BMC Medical Informatics and Decision Making 21, 69 (2021)

  @article{campillosetal-midm2021,   
  title       = {A clinical trials corpus annotated with UMLS entities to enhance the access to Evidence-Based Medicine},  
  author       = {Campillos-Llanos, Leonardo and Valverde-Mateos, Ana and Capllonch-Carri{\'o}n, Adri{\'a}n and Moreno-Sandoval, Antonio},   
  journal = {BMC Medical Informatics and Decision Making},
  volume = {21},
  number = {69},
  year      = {2021}
  }

medical-ner's Issues

Dataset provider

hi, I am trying to reproduce the reported results in your paper. From your code repo, I have not seen the dataset with BIO tags, is it possible to release the sentence level dataset recently?