Skip navigation
Please use this identifier to cite or link to this item:
Title: Écoute social et analyse des sentiments appliquée aux dialectes arabes et multilingues
Authors: ALAHOUM, NAdjia
KOUADRI, AIchouch RAnia
Keywords: Sentiment Analysis
Social Networks
Arabic Dialect
Deep Learning
Machine Learning
Issue Date: 2021
Abstract: Social networks are full of opinions about products and services that influence the buyin decisions of other consumers and affect the brand’s trustworthiness. In order to preserve their online identity, companies are looking for ways to analyze these opinions. Sentiment analysis, an application of natural language processing, can analyze comments and extract the opinion or sentiment behind them. Sentiment analysis is generally far from perfect ; Internet users, especially Algerians, do not make the task easy, they often express themselves with ”reinvented” spelling, from different languages and Arabic dialects simultaneously. Hence the difficulty of correctly categorizing each comment expressed. In this project, we are interested in Arabic dialects and multilingual sentiment analysis approaches. Our goal is to create a system that first performs social listening on twitter, and then, with the application of machine learning and deep learning, it automatically analyzes these collected tweets or comments whether they are written in Arabic and its dialects or in multilingual to finaly classify them according to their positive or negative polarity.
Description: Pr. Sidi Mohamed BENSLIMANE Encadreur Dr. Badia KLOUCHE Co-Encadreur
Appears in Collections:Ingénieur

Files in This Item:
File Description SizeFormat 
Final_m_moire__de_PFE (1).pdf286,29 kBAdobe PDFView/Open
Show full item record

Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.