CLÉMENT, Lionel; LANG, Bernard; SAGOT, Benoît

Métadonnées

Licence d’utilisation du document

CLÉMENT, Lionel
Laboratoire Bordelais de Recherche en Informatique [LaBRI]
Linguistic signs, grammar and meaning: computational logic for natural language [SIGNES]

LANG, Bernard
Software tools for natural language [ATOLL]

SAGOT, Benoît
Software tools for natural language [ATOLL]

Langue

Communication dans un congrès

Ce document a été publié dans

LREC 04, LREC 04, LREC 04, 2004, Lisbonne. 2004p. 1841-1844

Résumé en anglais

In this article, we introduce a new technique for constructing wide-coverage morphological lexica from large corpora and morphological knowledge, with an application to French. Basically, it relies on the idea that the existence of a hypothetical lemma can be guessed if several different words found in the corpus are best interpreted as morphological variants of this lemma. We first validated our technique by extracting verbs and adjectives on a general French corpus of 25 million words. Compared with other lexical resources available for French, our results are very satisfying, since we cover many words, often derived words, that are not always present in other lexica. Application of our algorithm to the acquisition of domain-specific adjectives on a botanic corpus gave also very good results, thus demonstrating its usability to extract domain-specific lexica. Moreover, it is generalizable to any language with a substantial morphology.< Réduire

Métadonnées

Partager cette publication !

Licence d’utilisation du document

Morphology based automatic acquisition of large-coverage lexica

Langue

Ce document a été publié dans

Résumé en anglais

URI

Origine

Unités de recherche