The application of network agenda setting model during the COVID-19 pandemic based on latent dirichlet allocation topic modeling.

Abstract

Based on Network Agenda Setting Model, this study collected 42,516 media reports from Party Media, commercial media, and We Media of China during the COVID-19 pandemic. We trained LDA models for topic clustering through unsupervised machine learning. Questionnaires (N = 470) and social network analysis methods were then applied to examine the correlation between media network agendas and public network agendas in terms of explicit and implicit topics. The study found that the media reports could be classified into 14 topics by the LDA topic modeling, and the three types of media presented homogeneity in the topics of their reports, yet had their own characteristics; there was a significant correlation between the media network agenda and the public network agenda, and the We Media reports had the most prominent effect on the public network agenda; the correlation between the media agenda and the implicit public agenda was higher than that of the explicit public agenda. Overall, findings showed a significant correlation between network agendas among different media.