University Sétif 1 FERHAT ABBAS Faculty of Sciences
Résultat de la recherche
1 résultat(s) recherche sur le mot-clé 'Hate speech Oensive language Videos Natural language processing'
Ajouter le résultat dans votre panier Affiner la recherche Générer le flux rss de la recherche
Partager le résultat de cette recherche
Titre : Machine and Deep Learning for Detection of Hate Speech in Videos Type de document : texte imprimé Auteurs : Kram,Amira, Auteur ; Toumi,Lyazid, Directeur de thèse Editeur : Setif:UFA Année de publication : 2021 Importance : 1 vol (67 f .) Format : 29 cm Langues : Français (fre) Catégories : Thèses & Mémoires:Informatique Mots-clés : Hate speech
Oensive language
Videos
Natural language
processingIndex. décimale : 003.54 Théorie de l'information Résumé :
In the last decade, the social network became popular, and the emergence
of the phenomenon of hate speech has been observed. A streaming platforms
like YouTube contains a lot of videos in dierent languages. The Arabic
videos with hate speech become a noticeable problem that requires the development
of automated tools to detect oensive language that aects all
categories of people who use YouTube.
The Arabic language is a Semitic language, but unfortunately, there is a
few scientic research concerning this language. The limited availability of
tools using the Arabic language makes to propose an automation tool is more
dicult. To our knowledge, this work is the rst that propose an automation
tool for detecting hate speech in Arabic videos.
In this master thesis, we propose to build an Arabic videos dataset from
the YouTube stream platform, how we annotated it, and using NLP techniques
for pre-processing step. Then, we applied popular machine learning
classiers using BOW, ngrams, TF-IDF and we propose deep learning methods
to solve our problem. Finally, the experiments on the used dataset show
that the support vector machines model gives the best performance for our
problem than the best known other classiers.
Côte titre : MAI/0475 En ligne : https://drive.google.com/file/d/1YUryxThPYFAP9g3mN0eCTHG4ZpUP8VT_/view?usp=shari [...] Format de la ressource électronique : Machine and Deep Learning for Detection of Hate Speech in Videos [texte imprimé] / Kram,Amira, Auteur ; Toumi,Lyazid, Directeur de thèse . - [S.l.] : Setif:UFA, 2021 . - 1 vol (67 f .) ; 29 cm.
Langues : Français (fre)
Catégories : Thèses & Mémoires:Informatique Mots-clés : Hate speech
Oensive language
Videos
Natural language
processingIndex. décimale : 003.54 Théorie de l'information Résumé :
In the last decade, the social network became popular, and the emergence
of the phenomenon of hate speech has been observed. A streaming platforms
like YouTube contains a lot of videos in dierent languages. The Arabic
videos with hate speech become a noticeable problem that requires the development
of automated tools to detect oensive language that aects all
categories of people who use YouTube.
The Arabic language is a Semitic language, but unfortunately, there is a
few scientic research concerning this language. The limited availability of
tools using the Arabic language makes to propose an automation tool is more
dicult. To our knowledge, this work is the rst that propose an automation
tool for detecting hate speech in Arabic videos.
In this master thesis, we propose to build an Arabic videos dataset from
the YouTube stream platform, how we annotated it, and using NLP techniques
for pre-processing step. Then, we applied popular machine learning
classiers using BOW, ngrams, TF-IDF and we propose deep learning methods
to solve our problem. Finally, the experiments on the used dataset show
that the support vector machines model gives the best performance for our
problem than the best known other classiers.
Côte titre : MAI/0475 En ligne : https://drive.google.com/file/d/1YUryxThPYFAP9g3mN0eCTHG4ZpUP8VT_/view?usp=shari [...] Format de la ressource électronique : Exemplaires (1)
Code-barres Cote Support Localisation Section Disponibilité MAI/0475 MAI/0475 Mémoire Bibliothéque des sciences Anglais Disponible
Disponible