Monitoring the Relationship between Social Network Status and Influenza Based on Social Media Data

Abstract

Abstract Background: This article aims to analyze the relationship between user characteristics on social networks and influenza. Methods: Three specific research questions are investigated: (1) we classify Weibo updates to recognize influenza-related information based on machine learning algorithms and propose a quantitative model for influenza susceptibility in social networks; (2) we adopt in-degree indicator from complex networks theory as social media status to verify its coefficient correlation with influenza susceptibility; (3) we also apply the LDA topic model to explore users’ physical condition from Weibo to further calculate its coefficient correlation with influenza susceptibility. From the perspective of social networking status, we analyze and extract influenza-related information from social media, with many advantages including efficiency, low cost, and real time. Results: We find a moderate negative correlation between the susceptibility of users to influenza and social network status, while there is a significant positive correlation between physical condition and susceptibility to influenza. Conclusions: Our findings reveal the laws behind the phenomenon of online disease transmission, and providing important evidence for analyzing, predicting, and preventing disease transmission. Also, this study provides theoretical and methodological underpinnings for further exploration and measurement of more factors associated with infection control and public health from social networks.