For effective management of urban extreme rainstorm disasters, promptly understanding public risk-perceived emotions of heavy rainfall is crucial to enhance governmental risk communication and emergency response. Therefore, this study develops a sentiment analysis framework that combines a rainstorm-specific sentiment lexicon with a deep learning model. By utilizing large-scale social media data, the framework further achieves dynamic monitoring and early warning of the public risk-perceived emotions during extreme rainstorm events. Specifically, this paper employed text mining techniques to analyze the emotional features of 51,222 Weibo posts related to rainstorm disasters, thereby constructing a specialized sentiment lexicon. Additionally, the lexicon was integrated with a TextCNN model to create a sentiment knowledge-enhanced hybrid sentiment analysis model. This hybrid method demonstrates an 11% increase in accuracy over the sole use of deep learning models. Moreover, an empirical analysis of the 2023 Zhuozhou extreme rainstorm disaster showed the framework's efficacy. Findings reveal that our method can effectively monitor the public risk-perceived emotions and provide early warning of risk anomalies during severe rainstorms by using the emotion index, which yields valuable insights for governmental bodies to accurately understand public risk perception and the dynamic evolution of disaster scenarios.
 

extreme rainstorm disasters; risk-perceived emotions; text mining techniques; sentiment lexicon; deep learning model