Linguistic study of Arabic lexical disambiguation with the aim of improving automatic translation

Abstract

The work presented in this thesis explores the question of the automatic acquisition of senses for lexical disambiguation in a translation framework. the issue of the identification of the senses is situated in a linguistic framework and the treatment is oriented towards translation. Thus, We propose a linguistic method of meaning acquisition and making it possible to establish semantic correspondences through creating a linguistic annotated corpora by taking advantage of the large amount of annotated English corpus Princeton WordNet sense, and using a machine translation systems. This method is applied within the Arabic language and evaluated on the Arabic corpus alone in a translation-linguistic framework.