Skip to Main content Skip to Navigation
Conference papers

Extraction terminologique : vers la minimisation de ressources

Yuliya Korenchuk 1, *
Abstract : The article presents the method which aims to minimize the use of external resources for the terminologyextraction task and to make this task less langage dependent. For that purpose, the method builds simplified morphologicaland morphosyntactic resources directly from a lemmatized corpus. These endogenous resources are used both in filters,which refine the statistical calculations, and in patterns for polylexical terms identification. The method was tested on twocomparable corpora in chemistry and in telecommunication in French and in English. The precision observed on the first100 monolexical terms fluctuates between 71% and 87% for French and between 44% and 69% in English ; for polylexicalterms the precision was 69-78% in French and 69-85% in English depending on the domain.
Complete list of metadatas

Cited literature [12 references]  Display  Hide  Download

https://hal.archives-ouvertes.fr/hal-01091663
Contributor : Yuliya Korenchuk <>
Submitted on : Friday, December 5, 2014 - 5:36:32 PM
Last modification on : Monday, January 20, 2020 - 3:26:02 PM
Long-term archiving on: : Monday, March 9, 2015 - 6:05:54 AM

File

Paper_O-L3.3.pdf
Files produced by the author(s)

Identifiers

  • HAL Id : hal-01091663, version 1

Collections

Citation

Yuliya Korenchuk. Extraction terminologique : vers la minimisation de ressources. TALN-RECITAL, Jul 2014, Marseille, France. pp.59-70. ⟨hal-01091663⟩

Share

Metrics

Record views

229

Files downloads

581