Revue de l'Information Scientifique et Technique
Volume 27, Numéro 2, Pages 24-28
2023-11-08
Authors : Bouchal Hakim . Belaid Ahror .
The containment caused by the coronavirus has stimulated the use of social networks as a means of exchange, communication, and combating social distancing. This paper describes our participation in the NLP Challenge 2022 competition initiated by "Centre de Recherche sur l’Information Scientifique et Technique" (CERIST) for the task of detecting hate speech in Arabic and offensive language on social networks. This task consists of analyzing Twitter messages related to the COVID-19 pandemic and classifying the users sentiment if they are hateful or not. In the present work, we propose a model based on recurrent neural networks, more precisely the Bidirectional long-term memory (Bi-LSTM), trained on a dataset built by the authors of this challenge. As result we achieves an accuracy of 96.35%.
Sentiment Analysis ; Offensive language detection ; Arabic Social Media ; Arabic text classification
Magnossão De Paula Angel Felipe
.
Bensalem Imene
.
Rosso Paolo
.
Zaghouani Wajdi
.
pages 7-12.
Chiker Nour Elhouda
.
pages 36-43.
Said Bachir
.
Barmati Mohammed Elsadiq
.
pages 13-17.
Geet D’sa Ashwin
.
Illina Irina
.
Fohr Dominique
.
pages 1-12.