I had to present a demo for Named Entity Recognition NER on Medical text Data. I found Stanza NLP Package suitable for my task but I was not able to integrate it with Spacy Displacy to show highlighted Entities because Spacy Model output and Stanza Model output for NER were very different. So I created this little script for the demo. I used Streamlit for UI.
I have used i2b2 clinical NER model trained on publicly available MIMIC-III database. It has been trained to extract following NER:
- PROBLEM
- TEST
- TREATMENT
For more information, you can visit stanza website.
I tested only in Python 3.7.13 in Linux.
- pip install -r requirements.txt
Execute the following statement - streamlit run ner.py
Open URL in browser.
Note: It will download some packages for the first time. So it might take some time to start. See terminal output for that.