Exploring the role of information security news descriptions on retweet proneness and user interactions

Bibliographic Data

ID7158771
AuthorsKonstantinos Charmanas (0000-0001-5743-609X, Aristotle University of Thessaloniki, corresponding author), Klairi Filippou (Aristotle University of Thessaloniki), Nikolaos Mittas (0000-0003-3061-7864, Democritus University of Thrace), Lefteris Angelis (0000-0002-6677-4039, Aristotle University of Thessaloniki)
Year2025
Volume8
Issue3
Publication date2025-08-01
Peer ReviewedYes
Open AccessYes
TypeARTICLE
VenueJournal of Computational Social Science (JOURNAL)
Journal identifiersISSN: 2432-2725 • E-ISSN: 2432-2717
PublisherSpringer Science and Business Media LLC (PUBLISHER)
DOI10.1007/s42001-025-00396-6
OpenAlexW4410867760
LanguageEN
References cited67

Nowadays, advanced network and resource capabilities offer many benefits to platform users but also bring malicious opportunities, thus leading experts to raise awareness about malicious threats and discuss potential mitigation practices through security news. An important factor in understanding user engagement and experiences is to investigate their sentiment and interests regarding security threats and tools across online platforms. In this study, we investigate content-based factors in security news affecting user interactions through a dataset of 44,264 tweets posted by seven relevant accounts. The first goal is to discover whether the textual information hidden in security news triggers retweeting through the training and evaluation of a set of classifiers. The findings suggest that words and hashtags can be important in developing prediction mechanisms. The second goal is to distinguish topics of security news leading to relatively more user interactions than the rest, where the topics are discovered using the Non-negative Matrix Factorization algorithm. For this goal, four types of user interactions are studied both independently and aggregated using the Archetypal Analysis and Conover-Iman test, respectively. The outcomes from these two approaches suggest that hacking activities followed by learning materials and webinars should be considered the most popular topics. Overall, the discussed findings can be used to understand the interests and reactiveness of Twitter users across security news, while the framework can be studied for extracting knowledge from Twitter data

Computer security · Data science · Internet privacy · World Wide Web · Complex Network Analysis Techniques · Computer Science · Misinformation and Its Impacts · Spam and Phishing Detection

Tools

Open DOIOpen Access