Application of Information Diffusion of Negative Emotion and Binomial Regression Method on Retweet Cascades during Covid-19

Abstract

This study based on text mining method on Weibo to reveal the relation of negative emotion and information cascade to the coronavirus (Covid-19) pandemic. With 176934 Weibo posts collected with the keyword “novel coronavirus pneumonia” from January 23 to February 7 2020, negative binomial regression method is applied to construct an empirical model between negative emotion and the cascade. Regression results demonstrated that emotional intensity is crucial to information cascade and the intensity threshold exist. The research also proved that positive is more effective than negative to the viral diffusion.