Relation Extraction from Clinical Cases for a Knowledge Graph
Extraction de relations dans des cas cliniques pour des graphes de connaissances
Résumé
We describe a system for automatic extraction of semantic relations between entities in a medical corpus of clinical cases. It builds upon a previously developed module for entity extraction and upon a morphosyntactic parser. It uses experimentally designed rules based on syntactic dependencies and trigger words, as well as on sequencing and nesting of entities of particular types. The results obtained on a small corpus are promising. Our larger perspective is transforming information extracted from medical texts into knowledge graphs.
Domaines
Informatique et langage [cs.CL]| Origine | Fichiers produits par l'(les) auteur(s) |
|---|