Multi-scale Sentiment Analysis of Location-Enriched COVID-19 Arabic Social Data

Tarek Elsaka, Imad Afyouni, Ibrahim Abaker Targio Hashem, Zaher AL-Aghbari

Research output: Chapter in Book/Report/Conference proceedingConference contribution

2 Citations (Scopus)

Abstract

After the recent outbreak of COVID-19, researchers have risen working on several challenges related to the mining of social data to learn about people’s reactions to the epidemic. Recent studies have largely focused on extracting current themes and inferring broad attitudes, with a particular emphasis on the English language. This study presents various perspective for Arabic social data mining to provide in-depth insights related to the COVID-19 pandemic. We initially devised a method for inferring geographical whereabouts from Arabic tweets not initially geotagged. Secondly, a sentiment analysis mechanism based on Arabic word embeddings is introduced, with several levels of geographical granularity (regions/countries) considered. Sentiment-based classifications of topics and subtopics related to COVID-19 will also be presented. According to our findings, the overall percentage of location-enabled tweets has increased from 2% to 46% (about 2.5M tweets). During the pandemic, Arab Twitter users’ negative emotions about lockdown, restriction, and law enforcement were also widely expressed.

Original languageEnglish
Title of host publicationDiscovery Science - 24th International Conference, DS 2021, Proceedings
EditorsCarlos Soares, Luis Torgo
PublisherSpringer Science and Business Media Deutschland GmbH
Pages194-203
Number of pages10
ISBN (Print)9783030889418
DOIs
Publication statusPublished - 2021
Externally publishedYes
Event24th International Conference on Discovery Science, DS 2021 - Virtual, Online
Duration: Oct 11 2021Oct 13 2021

Publication series

NameLecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Volume12986 LNAI
ISSN (Print)0302-9743
ISSN (Electronic)1611-3349

Conference

Conference24th International Conference on Discovery Science, DS 2021
CityVirtual, Online
Period10/11/2110/13/21

Keywords

  • Arabic COVID-19
  • Arabic tweets
  • COVID-19 pandemic
  • Sentiment Analysis
  • Social data mining

ASJC Scopus subject areas

  • Theoretical Computer Science
  • General Computer Science

Fingerprint

Dive into the research topics of 'Multi-scale Sentiment Analysis of Location-Enriched COVID-19 Arabic Social Data'. Together they form a unique fingerprint.

Cite this