Causal Graph and Social Network Analysis for the Spread of COVID-19 from Self-reported Indicator Data
Shaoou Chen, Yao Xie, Shihao Yang · 2021 55th Asilomar Conference on Signals, Systems, and Computers · 2021
Dynamic Bayesian Network (DBN) is an useful tool to learn the causal inference and social network of random variables. In this article, we analyze the correlations between the spread of coronavirus (COVID-19) and certain self-reported COVID-19 indicators in the United States, and then adopt DBN model with search and score-based approach to analyze and interpret the causal relationships and social network between these variables by learning the structure of the Directed Acyclic Graph from the model. We explore the change of causality among fifty states during the pandemic of COVID-19 in the year of 2020 and interpret the root cause for changes and trends. We concentrate on five worst states with COVID-19 and then extended our studies to all states by comparing the causal relationships and analyzing the patterns of DAG.