Institutional Repository
Technical University of Crete
Technical University of Crete
EN
|
EL
| URI: | http://purl.tuc.gr/dl/dias/531F2FFF-3CC0-49B2-B43D-A8643BAFAEFF | ||
| Year | 2023 | ||
| Type of Item | Diploma Work | ||
| License |
|
||
| Bibliographic Citation | Georgios-Nektarios Nikou, "An experimental analysis of Twitter suspension during the first COVID19 period", Diploma Work, School of Electrical and Computer Engineering, Technical University of Crete, Chania, Greece, 2023 https://doi.org/10.26233/heallink.tuc.97809 | ||
| Appears in Collections |
This study aims to capture the overall sentiment of people’s tweets regarding COVID-related subjects and to examine any attempts to spread fake news and misinformation on Twitter. Our research is based on a dataset collected through the Twitter API, containing approximately 200 million tweets from two popular COVID-related hashtags. We conduct sentiment analysis using the XLM-RoBERTa-large model on several topics related to the COVID-19 pandemic. Next, we perform data analysis to identify interesting patterns and characteristics of this vast dataset. Our research also targets suspended Twitter accounts and by using the Latent Dirichlet Allocation algorithm we identify their topics of discussion. We construct the retweet social graph to analyze their social network connections, enabling us to detect any coordinated actions to retweet the same content in large quantities. The results showed a trend in sentiment towards terms like COVID-19, conspiracy, and lockdown. We observe that although suspended users made up only 0.74% of the total users in the dataset, they generated 7.52% ofthe total posts in the dataset.
Copyright © DIAS 2013
