Recherche contextuelle d’équivalents en banque de terminologie

Download
  1. (PDF, 419 KB)
AuthorSearch for:
TypeArticle
Proceedings titleProceedings of the 17th Conférence sur le Traitement Automatique des Langues Naturelles (TALN 2010)
Conference17th Conférence sur le Traitement Automatique des Langues Naturelles (TALN 2010), July 19-23, 2010, Montreal, Québec
Pages# of pages: 10
Subjectcontextual search; term equivalents; term bank; domain-based disambiguation
AbstractOur research shows the usefulness of taking into account the context of a term within a text to be translated to better find an appropriate term equivalent for it in a term bank. A term bank has the particularity of categorising its records by assigning them one or more domains from a pre-established list of domains. The strategy presented here uses this domain information. An algorithm has been developed to automatically assign a domain profile to a source text. It is then combined with another algorithm which finds a match between a term’s domains (as found in the term bank) and the text’s domain profile. For our experimentation, bilingual abstracts (French-English) from eight scientific journals provide 1130 pairs of term equivalents. The Grand Dictionnaire Terminologique (Office Québécois de la Langue Française) is used as a terminological ressource. On our data set, we show a reduction of 75% in the average rank of the correct equivalent, in comparison to a random choice.
Publication date
LanguageEnglish
AffiliationNational Research Council Canada (NRC-CNRC); NRC Institute for Information Technology
Peer reviewedYes
NPARC number16067302
Export citationExport as RIS
Report a correctionReport a correction
Record identifier7ee401d0-dd1b-45b7-936b-31de38218075
Record created2010-11-03
Record modified2016-05-09
Bookmark and share
  • Share this page with Facebook (Opens in a new window)
  • Share this page with Twitter (Opens in a new window)
  • Share this page with Google+ (Opens in a new window)
  • Share this page with Delicious (Opens in a new window)