Detecting Social Media Rumor Debunking Effectiveness During Public Health Emergencies: An Interpretable Machine Learning Approach

Abstract

Debunking offers a promising approach to counteracting social media rumors during public health emergencies. However, the effective mechanisms of rumor debunking on social media remain unverified. This study employs an interpretable machine learning approach, combined with information and communication theories, to investigate social media rumor debunking effectiveness and its influencing factors. A total of 10,150 COVID-19 rumor-debunking posts and other relevant data on Sina Weibo were collected for analysis. The results showed that the beneficial impacts of debunking rumors surpass the adverse consequences and revealed significant differences in debunking effectiveness across diverse rumor types, topics, and involvement levels.