Arabic Text Classification Based on Word and Document Embeddings

Abdelkader El Mahdaouy 1 Eric Gaussier 2, 3, 1 Saïd El Alaoui Ouatik 4
1 AMA - Analyse de données, Modélisation et Apprentissage automatique [Grenoble]
LIG - Laboratoire d'Informatique de Grenoble
2 UGA - Université Grenoble Alpes
3 LIG - Laboratoire d'Informatique de Grenoble
4 LIM - laboratoire informatique et modélisation
Abstract : Recently, Word Embeddings have been introduced as a major breakthrough in Natural Language Processing (NLP) to learn viable representation of linguistic items based on contextual information or/and word co-occurrence. In this paper, we investigate Arabic document classification using Word and document Embeddings as representational basis rather than relying on text preprocessing and bag-of-words representation. We demonstrate that document Embeddings outperform text preprocessing techniques either by learning them using Doc2Vec or averaging word vectors using a simple method for document Embedding construction. Moreover, the results show that the classification accuracy is less sensitive to word and document vectors learning parameters.
Keywords : Doc2vec Arabic text classification Arabic natural language processing Document embeddings Word embeddings SKIP-Gram Continuous Bag-of-Word Glove
Document type :
Book sections
Liste complète des métadonnées
https://hal.archives-ouvertes.fr/hal-01386136
Contributor : Abdelkader El Mahdaouy <>
Submitted on : Saturday, October 22, 2016 - 8:30:41 PM
Last modification on : Thursday, October 11, 2018 - 8:48:05 AM

Identifiers

Collections

LIG_SRCPR | LIG | UGA

Citation

Abdelkader El Mahdaouy, Eric Gaussier, Saïd El Alaoui Ouatik. Arabic Text Classification Based on Word and Document Embeddings. Advances in Intelligent Systems and Computing , 533, Springer International Publishing, pp 32-41, 2016, Proceedings of the International Conference on Advanced Intelligent Systems and Informatics 2016, 978-3-319-48307-8. 〈10.1007/978-3-319-48308-5_4〉. 〈hal-01386136〉

Export

BibTeX TEI DC DCterms EndNote

Share

Metrics

Record views

316

Contact

support.ccsd.cnrs.fr
hal.support@ccsd.cnrs.fr
Logo CCSD