MEANS: a semantic approach to the search for answers to medical questions

We present a medical question answering approach, called MEANS. This approach relies on natural language processing techniques to extract medical entities and relations from the user questions and medical corpora. It also uses semantic Web languages to represent and query the information searched by...

Full description

Saved in:
Bibliographic Details
Published in:Traitement Automatique des Langues 2014-01, Vol.55 (1), p.71-104
Main Authors: Abacha, Asma Ben, Zweigenbaum, Pierre
Format: Article
Language:fre
Online Access:Get full text
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:We present a medical question answering approach, called MEANS. This approach relies on natural language processing techniques to extract medical entities and relations from the user questions and medical corpora. It also uses semantic Web languages to represent and query the information searched by the users. This feature allows to share the information extracted from textual corpora using standard languages and to consider incremental knowledge acquisition in mid-and-long terms. MEANS constructs an initial SPARQL query and several relaxed queries as semantic interpretations of a user question. The evaluation of MEANS on a real dataset shows promising results for both precision and MRR and showed the significant benefits of the query relaxation technique. Adapted from the source document
ISSN:1965-0906