ncp's Introduction

NCP : NLP Cancer Pipeline

NCP: a module responsible for converting clinical judgments written in spanish natural language into a JSON format with the most important fields for analysis and data mining.

Diagrams

Data Flow

Technology

Steps

Preprocessing

Acronyms are only expanded if they are known link and do not conflict with other words. The main objective is to correct spelling errors to improve network results.

Annotation

This is done using the Prodigy tool from Explosion.

Transformers Networks

NER models, fine-tuned from a pre-trained model.

Two models:

Postprocessing:

Two phases:

Normalize

Definitive terms from the database.

CMDS

To test it

python test.py

Build package

python setup.py sdist

see Makefile file

Citation

If you use our work in your research, please cite it as follows:

@INPROCEEDINGS{garcia-barraganCBMS2023,
  author={García-Barragán, Alvaro and Solarte-Pabón, Oswaldo and Nedostup, Georgiy and Provencio, Mariano and Menasalvas, Ernestina and Robles, Victor},
  booktitle={2023 IEEE 36th International Symposium on Computer-Based Medical Systems (CBMS)},
  title={Structuring Breast Cancer Spanish Electronic Health Records Using Deep Learning},
  year={2023},
  pages={404-409},
  keywords={Natural Language Processing (NLP), Information extraction, Deep Learning, Breast cancer.},
  doi={10.1109/CBMS58004.2023.00252}
}

Recommend Projects