A neural approach for sentiment analysis in Algerian dialect

Notice brève
dc.contributor.authorHAMADOUCHE Khaoula
dc.date.accessioned2025-07-08T10:21:05Z
dc.date.available2025-07-08T10:21:05Z
dc.date.issued2025-05-29
dc.description.abstractOur research explores several fundamental aspects of natural language processing for sentiment analysis, with a focus on automatic corpus construction and sarcasm detection within the context of Algerian dialect.Specifically, we have adopted state-of-the-art techniques in deep learning, namely the BERT model (Bidirectional- Encoder Representations from Transformers). This research aims to address key challengesrelatedtounderstandingandprocessing linguistic nuances in Algerian Arabictextsbydesigningaplatformthat collects data from various social media sources and conducts sentiment analysistasks(polaritydeterminationandsarcasm detection). Methodologically,ourthesisemploys techniques including: (1) corpus extraction from social media, (2) the application of machine learning algorithms for sentiment classification, and (3) the integration of linguistic rules and deep learning models for sarcasm detection. The approaches we have used have allowed us to achieve satisfactory results that can extend to various applications in social media analysis, opinion extraction, and automated content analysis within arabophone communities, including the study of dialectal variations. This thesis underscores the importance of linguistic diversity in NLP researchand provides a foundation for future investigations in the field of dialectal sentiment analysis.
dc.formatPDF
dc.identifier.urihttps://dspace.univ-oran1.dz/handle/123456789/4653
dc.language.isoen
dc.publisherUniversité oran1
dc.subjectNatural Language Processing ; Sentiment Analysis ; Algerian Dialect ; Corpus Construction ; BERT ; sarcasm; deep learning; machine learning; Arabic natural processing ; WORD emberedding
dc.titleA neural approach for sentiment analysis in Algerian dialect
dc.typeThesis
grade.ExaminateurHAMDADOU Djamila, Professeur, Université Oran 1
grade.ExaminateurAMAR BENSABER Djamel, Professeur, Ecole Supérieure en Informatique Sidi Bel Abbes
grade.ExaminateurMOUSSAOUI Abdelouahab, Professeur, Université Sétif 1 Ferhat Abbas
grade.ExaminateurMEZIANE Hassina, MCA, Université Oran 1
grade.GradeDoctorat 3éme cycle
grade.OptionSystèmes informatiques et réseaux
grade.PrésidentABDI Mustapha Kamel, Professeur, Université Oran 1
grade.RapporteurBOUSMAHA Kheira Zineb, MCA, Université Oran 1
l'article.1.DateParution19 Decembre 2023
l'article.1.RevueACM Transactions on Asian and Low-Resource Language Information Processing
l'article.1.RéférenceKhaoulaHamadouche, KheiraZinebBousmaha, Mohamed AbdelwaretBekkoucha, and Lamia Hadrich-Belguith. 2023. AlgBERT: Automatic Construction of Annotated Corpus for Sentiment Analysis in Algerian Dialect. ACM Trans. Asian Low-Resour. Lang. Inf. Process. 22, 12, Article 257 (December 2023), 17 pages. https://doi.org/10.1145/3632948
l'article.1.TitreAlgBERT: Automatic Construction of Annotated Corpus for Sentiment Analysis in Algerian Dialect
l'article.2.DateParution31 Decembre 2022
l'article.2.RevueRevue d'Intelligence Artificielle
l'article.2.RéférenceBousmaha, K.Z., Hamadouche, K., Gourara, I., Hadrich, L.B. (2022). DZ-OPINION: Algerian dialect opinion analysis model with deep learning techniques. Revue d'Intelligence Artificielle, Vol. 36, No. 6, pp. 897-903. https://doi.org/10.18280/ria.360610
l'article.2.TitreDZ-OPINION: Algerian Dialect Opinion Analysis Model with Deep Learning Techniques
l'article.3.DateParution19 juillet 2024
l'article.3.RevueACM Transactions on Asian and Low-Resource Language Information Processing
l'article.3.RéférenceKheiraZinebBousmaha, KhaoulaHamadouche, HadjerDjouabi, and Lamia Hadrich-Belguith. 2024. Automatic Algerian Sarcasm Detection from Texts and Images. ACM Trans. Asian Low-Resour. Lang. Inf. Process. 23, 7, Article 108 (July 2024), 25 pages. https://doi.org/10.1145/3670403
l'article.3.TitreAutomatic Algerian Sarcasm Detection from Texts and Images
la.MentionTrès Honorables
la.SpécialitéInformatique
la.coteTH5595
Fichiers
Bundle original
Voici les éléments 1 - 1 sur 1
Vignette d'image
Nom :
TH5595.pdf
Taille :
7.13 MB
Format :
Adobe Portable Document Format
Description :
Télécharger
Collections
Thèses de Doctorat "Informatique"