Skip to Main content Skip to Navigation
Master Thesis

Prétraitement de données et création d'un segmenteur de l'arabe pour un système de traduction probabiliste vers le français

Abstract : In recent years, Natural Language Processing has rapidly evolved, especially in the domain of Statistical Machine Translation causing the need for reliable automatic translations to skyrocket with no sign of slowing. Due to this increased need, we have taken a special interest in this domain with the goal of creating a translation machine capable of translating Arabic into French, based on statistical models. The performance of Statistical Machine Translation relies heavily on the quality and the quantity of available training data. However, the Arabic language remains one of the languages with the fewest available resources which is why most of the available works in this language are based on open access data from international organizations, such as the U.N. In this work, we will present our approach to optimizing the performance quality of our Arabic translator. Taking into account the lack of data and available resources, we were able to find a low-cost solution to search for the best pre-processing combinations to apply to our Arabic database in order to obtain the highest quality French translation.
Document type :
Master Thesis
Complete list of metadata

Cited literature [44 references]  Display  Hide  Download

https://dumas.ccsd.cnrs.fr/dumas-00757706
Contributor : Uga - Bulles <>
Submitted on : Tuesday, November 27, 2012 - 2:29:07 PM
Last modification on : Tuesday, May 11, 2021 - 11:36:17 AM
Long-term archiving on: : Saturday, December 17, 2016 - 3:21:46 PM

Identifiers

  • HAL Id : dumas-00757706, version 1

Citation

Sahnoun Ben Taamallah. Prétraitement de données et création d'un segmenteur de l'arabe pour un système de traduction probabiliste vers le français. Sciences de l'Homme et Société. 2012. ⟨dumas-00757706⟩

Share

Metrics

Record views

447

Files downloads

1327