Informational approach to uncover the age group interactions in epidemic spreading from macro analysis.

Abstract

In this study, we explore transfer entropy (TE) as a tool to explore the evolution of population contact patterns in epidemic processes. Initially, we apply TE to a classical age-stratified SIR model and find that the inferred patterns align with the interaction structure of the population, as defined by the age-mixing matrix. Applying this methodology to the COVID-19 pandemic data from Spain, we illustrate how TE can capture temporal changes in individual behavior. Furthermore, we demonstrate that incorporating the inherent dynamics of the epidemic process allows us to create a coarse-grained representation of the time series, providing richer information than raw data. We argue that this macro-level perspective is enhanced by the effectiveness of causal analysis across different scales. Our findings underscore the potential of informational approaches to retrospectively track behavioral adaptations during a pandemic, offering valuable insights for tailoring strategies to control disease spread. This paper paves the way for future research into using such methods for model-free estimation of contact patterns during pandemics.