Research on earthquake emergency response modes of individuals based on social surveillance video

Abstract After the earthquake, the correct and efficient earthquake emergency response modes of individuals are crucial for reducing casualties. Based on the keyword database of the earthquake emergency response behaviors of individuals, this research used the machine learning method to construct a decision tree of the earthquake emergency response modes. Through the decision tree, this study extracted decision rules that can predict the emergency response decisions of individuals in future earthquake and also determined the relative importance of six different factors (Seismic intensity > Social context >Location>Action>Age>Gender) in determining response behaviors. The set of rules was quite effective when evaluated using the test individuals with an overall accuracy of 85.23%. And then, combined with the cultural background in China, we had extracted and analyzed the main response modes of individuals at seismic intensities I-V, VI, VII, VIII-IX stricken regions by earthquake. The decision rules and main response modes of individuals were scientific bases for the targeted earthquake emergency safety education and the government emergency rescue.