The goals of emergency management are to restore human safety and security, and to help the authorities understand what causes such events. It requires information that is both highly accurate, and can be generated very quickly. This research addresses these concerns with a machine learning model based on cause-and-effect using a Bayesian belief network. This employs human critical thinking and amplified context to encode the model structures, which contribute towards its imitation of human-intelligent understanding, and the model parameters are fitted using social media data. The results show that our model is a natural fit for a real-world environment required emergency management.

应急管理的目标是恢复人类安全和保障，并帮助当局了解导致此类事件的原因。它需要高度准确的信息，并且可以非常快速地生成。这项研究使用基于因果关系的机器学习模型使用贝叶斯信念网络解决了这些问题。这采用人类批判性思维和放大的上下文来编码模型结构，这有助于其模仿人类智能理解，并且使用社交媒体数据拟合模型参数。结果表明，我们的模型非常适合需要应急管理的真实环境。