Biomedical Entity Recognition in Bahasa Indonesia

Summary:

• Trained using manually annotated data from alodokter.com (online health QA platform) using UMLS guideline (see https://rdcu.be/cNxV3)
• Recognize disorders (DISO) and anatomy (ANAT) entities
• Achieve best F1 macro score 0.81
• Based on XLM-Roberta. So, cross lingual recognition might work.

CITATION

This work is done with generous support from Safitri Juanita, Dr. Diana Purwitasari and Dr. Mauridhi Hery Purnomo from Institut Teknologi Sepuluh Nopember, Indonesia.

Citation for academic purpose will be provided later.
In the meantime, please let me know whenever you use this model (mail to : abid(dot)famasya(at)gmail.com) :)

For demo, please go to the HF space demo: https://huggingface.co/spaces/abid/id-bioner-demo