Guidelines for Annotation of Arabic Dialectness - Archive ouverte HAL Accéder directement au contenu
Communication Dans Un Congrès Année : 2008

Guidelines for Annotation of Arabic Dialectness

Nizar Habash
  • Fonction : Auteur
  • PersonId : 836109
Owen Rambow
  • Fonction : Auteur
  • PersonId : 907836
Reem Kanjawi-Faraj
  • Fonction : Auteur
  • PersonId : 1361937

Résumé

The Arabic language is a collection of variants with phonological, morphological, lexical, and syntactic differences comparable to those among the Romance languages. However, throughout the Arab world, the standard written language is the same, Modern Standard Arabic (MSA). MSA is based on Classical Arabic and is itself not a native spoken language. MSA is used in some official spoken communication, such as newscasts, parliamentary debates, etc. Other forms of Arabic (generally referred to as "dialects" of MSA) are what people use for daily spoken communication. In this paper, we focus on the issue of creating standard annotation guidelines identifying dialect switching between MSA and at least one dialect in written text. These guidelines can form the basis for the annotation of large collections of data that will be used for training and testing automatic approaches to dialect identification and automatic processing of Arabic which exhibits dialect switching. We report on some initial annotation experiments: we discuss statistical distributions of labels in a small corpus (~19K words) that we annotated according to the guidelines and present inter-annotator agreement results.
Fichier principal
Vignette du fichier
57.pdf (235.04 Ko) Télécharger le fichier
Origine Fichiers produits par l'(les) auteur(s)

Dates et versions

hal-04497104 , version 1 (09-03-2024)

Licence

Domaine public

Identifiants

  • HAL Id : hal-04497104 , version 1

Citer

Nizar Habash, Owen Rambow, Mona Diab, Reem Kanjawi-Faraj. Guidelines for Annotation of Arabic Dialectness. Workshop on Arabic and its local languages, LREC, May 2008, Marrakech, Morocco. ⟨hal-04497104⟩
8 Consultations
8 Téléchargements

Partager

Gmail Mastodon Facebook X LinkedIn More