How to let gatortron-base identify medical terms from the text
#8
by
dgdhairya
- opened
I have extracted text from a pdf and I want to identify medical terms like BP, Hg, LDL, etc. For this purpose, I thought of using gatortron-base but I am not able to get any help as it needs data to be fine-tuned but I don't have much data to fine-tune the model. I thought of using it for NER purpose but still no use, does anyone have any idea how I can use gatortron itself for my work. If not gatortron, I am open to all suggestions.
Yes, you need to find your own dataset to fine-tune GatorTron for extraction, here is an example: https://github.com/uf-hobi-informatics-lab/ClinicalTransformerNER
We do not have the data for fine turning, you can find some benchmark datasets to fine-tune, such as the i2b2 2010 dataset.