Conference Papers Year : 2022

Sentiment Analysis on COVID-19 Twitter Data: A Sentiment Timeline

Makrina Karagkiozidou
  • Function : Author
  • PersonId : 1406718
Paraskevas Koukaras
  • Function : Author
  • PersonId : 1406719
Christos Tjortjis
  • Function : Author
  • PersonId : 1012048

Abstract

COVID-19 has been one of the most dominant discussion topics on Twitter since 2019. Users express their opinions representing public sentiment on the topic. This paper presents a sentiment timeline of Twitter users, regarding COVID-19 vaccines. This work raises concerns about the extracted information with regards to sentiment analysis, the dominance of each sentiment and its influential power. During the implementation of the analysis, several datasets were examined for the creation of the model. Various algorithms were employed with Random Forest performing best and therefore selected for training the model, achieving an accuracy of 91.5%. Our findings indicate that the majority of Twitter users are positive regarding COVID-19 vaccines and support WHO’s recommendations. Negative tweets comprising the minority of the tweets, appear to have a higher influential power with their retweet rates, outperforming positive and neutral sentiments.
Fichier principal
Vignette du fichier
534967_1_En_29_Chapter.pdf (646.91 Ko) Télécharger le fichier
Origin Files produced by the author(s)

Dates and versions

hal-04668644 , version 1 (07-08-2024)

Licence

Identifiers

Cite

Makrina Karagkiozidou, Paraskevas Koukaras, Christos Tjortjis. Sentiment Analysis on COVID-19 Twitter Data: A Sentiment Timeline. 18th IFIP International Conference on Artificial Intelligence Applications and Innovations (AIAI), Jun 2022, Hersonissos, Greece. pp.350-359, ⟨10.1007/978-3-031-08337-2_29⟩. ⟨hal-04668644⟩
23 View
12 Download

Altmetric

Share

More