CliniRes: Publicly Available Mapping of Clinical Lexical Resources

Abstract

This paper presents a human-readable resource for mapping identifiers from various clinical knowledge bases. This resource is a version of UMLS Metathesaurus enriched with WordNet 3.0 and 3.1 synsets, Wikidata items with their clinical identifiers, SNOMED CT to ICD-10 mapping and Spanish ICD-10 codes description. The main goal of the presented resource is to provide semantic interoperability across the clinical concepts from various knowledge bases and facilitate its integration into mapping tools. As a side effect, the mapping enriches already annotated medical corpora for entity recognition or entity linking tasks with new labels. We experiment with entity linking task, using a corpus annotated both manually and with the mapping method and demonstrate that a semi-automatic way of annotation may be used to create new labels. The resource is available in English and Spanish, although all languages of UMLS may be extracted. The new lexical resource is publicly available.

Resources

Details