Extracting Useful Emergency Information from Social Media: A Method Integrating Machine Learning and Rule-Based Classification

User-generated contents on social media are a valuable source of emergency information that can facilitate emergency responses. However, the tremendous amount and heterogeneous quality of social media data make it difficult to extract truly useful emergency information, especially using pure machine learning methods. This study proposes a machine learning and rule-based integration method (MRIM) and evaluates its emergency information classification performance and determinants. Using microblog data about the July 20 heavy rainstorm in Zhengzhou posted on China’s largest social media platform, the study finds that MRIM performs better than pure machine learning methods and pure rule-based methods. Its performance is influenced by microblog characteristics such as number of words, presence of exact address and contact information, and users’ attention. The research demonstrates the feasibility of integrating machine learning and rule-based methods to mine social media texts and provides actionable suggestions for emergency information management practitioners.

Hongzhou Shen, Yue Ju, Zhijing Zhu