Final Report for Social Media Misinformation: Detection and Impacts

Abstract

There is a large and growing body of well-documented social media misinformation related to the origin, treatment, and severity of the COVID19 epidemic. This research examines the impact of this misinformation on human health outcomes for several countries affected by COVID-19 at the sub-national level. This report responds to the investigation done to discover the location of misinformation networks to COVID19 case rates outside of the U.S. In addition, the message content collected has enabled the development of supervised machine learning classification algorithms to identify misinformation and include bursts contrary to public health authority advice. Burst detection methods are developed to capture spikes in the misinformation spreading related to COVID-19. The network theoretic approach was taken to discover the network of misinformation at finer spatial scales. Finally, the spatial correlation of social network exploitation and disease prevalence at the sub-national level was studied.