Can Network Public Opinion Crises Following Natural Disasters Be Prevented? Insight From an Emotion‐ and Information‐Driven Risk Early Warning Model Using Flood Events in China

This study aims to identify, assess, and warn about the network public opinion crises triggered by sudden natural disasters from an emotion‐ and information‐driven perspective. By integrating Information Ecology Theory and Emotions as Social Information Theory, the study develops a comprehensive early warning system for public opinion risks. The proposed model, combining Fuzzy Interpretive Structural Modelling and Bayesian Belief Network, demonstrates strong quantitative performance, achieving an accuracy of 92.16% in forecasting network public opinion crises under uncertain conditions. Real‐world case analysis validates its effectiveness, while sensitivity analysis identifies key factors—such as action tendencies, risk perceptions, user engagement, emotion diffusion, and emotion divergence—that significantly influence public opinion risk. Scenario simulations highlight that moderate government intervention, especially at early stages, optimally mitigates risks. The study provides a data‐driven framework that enhances risk assessment accuracy by considering interdependencies between risk factors and integrating fuzzy logic. Targeted insights can be drawn regarding the governance and mitigation of public opinion crises amid flood disasters within China's institutional and social context.