Multilingual Extraction of functional relations between Arabic Named Entities using NooJ platform

Abdelmajid Ben Hamadou (1) , Odile Piton (2) , Héla Fehri (1)

1 MIRACL - Multimedia, InfoRmation systems and Advanced Computing Laboratory
2 SAMM - Statistique, Analyse et Modélisation Multidisciplinaire (SAmos-Marin Mersenne)

The extraction of relation between Named Entities (NE) has become the last few years an interesting research domain. It is very useful for many applications such as Web mining, Information extraction and retrieval, Business intelligence, Automatic databases filing with Entities & types, Questions answering task and document Summarization. Several works has been performed for relation discovery in texts written in Latin languages and as far as we know, very few works has been done for Arabic language. In this paper, we focus on functional relations between ENAMEX and ORG Arabic Named Entities. The extraction approach is rule based and the implementation is performed using NooJ Platform.

Mots clés: Relations between Named Entities; Extraction Process; NooJ transducers; Relation translation

Domaines: Informatique et langage [cs.CL]

Abdelmajid Ben Hamadou, Odile Piton, Héla Fehri. Multilingual Extraction of functional relations between Arabic Named Entities using NooJ platform. Nooj 2010 International Conference and Workshop, Le Département de Philologie Grecque de l'Université Democritus de Thrace, le Laboratoire de Sémio-Linguistique et Didactique (LASELDI) de l'Université de Franche-Comté et la Maison des Sciences de l'Homme et de l'Environnement Ledoux, May 2010, Komotini, Greece. pp.192-202. ⟨hal-00547940⟩