In this month’s research spotlight, we highlight recent research from COSMOS that focuses on using epidemiological modeling to study information diffusion—specifically, the studies titled,

• “Are Narratives Contagious? Modeling Narrative Diffusion Using Epidemiological Theories,” and
• “Utilizing Fractional Order Epidemiological Model to Understand High and Moderate Toxicity Spread on Social Media Platforms.”
  

Each studied information diffusion, such as online toxicity spread, from the point of view of contagion using epidemiological models. These studies were published and presented recently at this year’s annual International Conference on Advances in Social Networks Analysis and Mining (ASONAM 2024) in early September, which took place from 2 to 5 September in Calabria, Italy.

“Are Narratives Contagious? Modeling Narrative Diffusion Using Epidemiological Theories,” examined how narratives spread on social media, particularly YouTube, by applying epidemiological models (SIR and SEIZ) to study narrative diffusion in the South China Sea dispute context. The study found that narratives spread like contagions, with the SEIZ model outperforming SIR in tracking narrative spread. Analysis of two main narratives revealed that content portraying the US as an aggressor was more “infectious” than Russia-Ukraine conflict content.

“Utilizing Fractional Order Epidemiological Model to Understand High and Moderate Toxicity Spread on Social Media Platforms,” introduced a fractional-order epidemiological model (SEI_mI_hQR) to study toxic content spread on social media platforms. This model differentiates between moderate and highly toxic users (I_m and I_h, respectively) and includes a quarantine intervention strategy. Using Twitter data of posts with the #Fcovid hashtag, the study found that quarantining the most toxic and active users resulted in very low error rates (0.0011–0.0012). Their model helps understand and predict toxic content spread better than conventional models.