UMLS-EDA

🎉 A light-weighted UMLS-based data augmentation for biomedical NLP tasks including Named Entity Recognition and sentence classification 🎉

• Citation: Kang, T., Perotte, A., Tang, Y., Ta, C., & Weng, C. (2020). UMLS-based data augmentation for natural language processing of clinical research literature. Journal of the American Medical Informatics Association.
• Author: Tian Kang (tk2624@cumc.columbia.edu)
• Affiliation: Department of Biomedical Informatics, Columbia Univerisity (Dr. Chunhua Weng's lab)
• Built upon EDA-Easy Data Augmentation

User Guide

0. Before start

1. Install 'UMLS' and 'QuickUMLS' locally
2. Get your UMLS SOAP API Key from the UTS ‘My Profile’ area after signing in UMLS Terminology service
3. Add your API Key and QuickUMLS directory to the config.py.
4. Costomzie other variables in the config.py

1. Named Entity Recognition

• Input: CoNLL format file
• Usage:

    python augment4ner.py [-h] --input INPUT [--output OUTPUT] [--num_aug NUM_AUG] [--alpha ALPHA]

2. Sentence Classification

• Input: "|" seperated file (index|label|sentence text)
• Usage:

    python augment4class.py [-h] --input INPUT [--output OUTPUT] [--num_aug NUM_AUG] [--alpha ALPHA]

See examples/example4ner.conll and example/example4class.txt

Reference

• Wei, J. and Zou, K., 2019. Eda: Easy data augmentation techniques for boosting performance on text classification tasks. arXiv preprint arXiv:1901.11196. (Github repo)