Epidemic proximity and imitation dynamics drive infodemic waves during the COVID-19 pandemic.

The spreading of COVID-19 pandemic is accompanied an by an infodemic, defined as an overabundance of misinformation and disinformation. Infodemic is an emergent phenomenon due to information dissemination with unintended consequences of the complex dynamics of human behavior.
Modeling infodemic waves which spread through online social media during an ongoing epidemic is a challenging problem. To tackle this challenge, we collected more than 180 millions Twitter messages posted across 40 countries worldwide, with the aim to analyze the infodemic dynamics and its causal relationships with epidemic outbreaks. We considered three different modeling setups, namely, a compartmental model inspired by epidemic processes (SIS), an evolutionary game theoretic model driven by imitation dynamics (DI), and a bounded rationality model that assumes not pure rational individual behaviours (BR). These three models are statistically compared among them in order to evaluate which one better explains the infodemic modulation around the beginning of local epidemic.
We found that the most effective model in both the reproduction of local infodemic waves and for out-of-sample predictions is an evolutionary model of population dynamics, driven by epidemic incidence as an exogenous signal. Our model advances our current understanding of the epidemic-infodemic interplay and their co-evolution dynamics, demonstrating that the emergence of infodemic in the digital sphere cannot be easily understood in terms of epidemic-like spreading.

Συνεδρία: 
Authors: 
Valeria D'Andrea, Oriol Artime, Nicola Castaldo, Riccardo Gallotti and Manlio De Domenico
Room: 
2
Date: 
Monday, December 7, 2020 - 15:00 to 15:15

Partners

Twitter

Facebook

Contact

For information please contact :
ccs2020conf@gmail.com