Modeling Protective Action Decision-making In Earthquakes By Using Explainable Machine Learning And Video Data

This article discusses the potential of machine learning in modeling human behavior during disasters, specifically in earthquake protective action decision-making. The authors present a methodological framework for identifying behavioral insights using data collected from CCTV footage and social media. The framework includes key factors such as environmental and social cues, as well as behavior states of decision-makers. The results can be valuable for stakeholders in designing effective solutions for disaster management.