Early Detection and Control of the Next Epidemic Wave Using Health Communications: Development of an Artificial Intelligence-Based Tool and Its Validation on COVID-19 Data from the US

Abstract

Social media networks highly influence on a broad range of global social life, especially in the context of a pandemic. We developed a mathematical model with a computational tool, called EMIT (Epidemic and Media Impact Tool), to detect and control pandemic waves, using mainly topics of relevance on social media networks and pandemic spread. Using EMIT, we analyzed health-related communications on social media networks for early prediction, detection, and control of an outbreak. EMIT is an artificial intelligence-based tool supporting health communication and policy makers decisions. Thus, EMIT, based on historical data, social media trends and disease spread, offers an predictive estimation of the influence of public health interventions such as social media-based communication campaigns. We have validated the EMIT mathematical model on real world data combining COVID-19 pandemic data in the US and social media data from Twitter. EMIT demonstrated a high level of performance in predicting the next epidemiological wave (AUC = 0.909, F1 = 0.899).

Using an epidemiological model to explore the interplay between sharing and advertising in viral videos

“…The phenomenon of viral videos has caused the shift of budgets of companies from mass media to online marketing activities 2 . Understanding how information goes viral has attracted attention in fields of psychology 3 , marketing 4,5 and public health management 6 . Insights of the spread dynamics of viral videos would be useful to many parties such as political campaigners and product marketing managers.…”

“…An autonomous system corresponding to (1) can be obtained by introducing some new variables Substituting these variables into (1) leads to Calculating the equilibria of ( 6) leads to a third-order polynomial for i, which suggests that the dynamics of ( 6) could be quite complicated. Furthermore, reaching a steady state in system (6) suggests that the proportion of individuals in each class holds constant as the spread of a video, while a consistently growing number of active sharers seems not realistic for the long-term dynamics of information dissemination. We leave further modifications and analysis as future work.…”

Corruption dynamics: a mathematical model and analysis

“…Globally, a great deal of research has been conducted on how pandemics evolve. The evaluation and forecasting of the corruption pandemic's spread, together with the identification of its parameters, were thoroughly and meticulously examined in the studies [9][10][11][12][13][14][15]. A mathematical model of corruption dynamics including stability analysis was created by Nathan and Jackob [13].…”