Repérages lexicométriques des équivalences à basse fréquence dans les corpus bilingues

Abstract : The approach suggested in this article enables statistic identification of low-frequency word correspondences of bilingual texts aligned on phrase level. Corresponding lexical units are discovered through characteristic element computation in parallel contexts. An extensive description of translation equivalence is obtained through the study of multiple co-occurrences. The calculation undergoes systematic reiteration in order to embrace the entire corpus. The exploratory results show that the use of quantitative methods in combination with a bilingual lexicon or a dictionary offers new prospects for improving automatic word alignment.
Document type :
Journal articles
Complete list of metadatas

https://hal-univ-diderot.archives-ouvertes.fr/hal-01224061
Contributor : Maria Zimina <>
Submitted on : Tuesday, November 3, 2015 - 10:30:04 PM
Last modification on : Friday, January 4, 2019 - 5:33:30 PM

Identifiers

  • HAL Id : hal-01224061, version 1

Collections

Citation

Maria Zimina. Repérages lexicométriques des équivalences à basse fréquence dans les corpus bilingues. Lexicometrica, André Salem, Serge Fleury, 2002, Numéro spécial "Alignement lexical dans les corpus multilingues". ⟨hal-01224061⟩

Share

Metrics

Record views

131