SEED (Symptom Extraction in English using Deep learning) is a pipeline for classifying social media texts that contain mentions of symptoms, extracting the spans of symptom mentions and normalizing them to UMLS CUIs for use in Public Health.

 

Link to preprint: https://www.medrxiv.org/content/10.1101/2021.02.09.21251454v2.full

SEED is available in a standalone version and as a REST API.

Code on Bitbucket: https://bitbucket.org/pennhlp/seed/src
[Code will be made available on publication of the article]