Assessment of Community Vulnerability to Natural Disasters in Korea by Using GIS and Machine Learning Techniques

Despite similar natural hazard magnitudes, the economic losses and fatalities due to natural disasters are usually unevenly distributed among nations, regions, communities, and individuals. Socially, economically, and environmentally vulnerable communities are more likely to suffer disproportionately from disasters. Identifying vulnerability factors to disasters is critical information for disaster managers and planners to make disaster-related policy and strategies for mitigating the negative impacts of disasters. This study constructs an index of disaster vulnerability of local communities in Korea. Twelve indicators including social, economic, and natural environment and built environment aspects are selected to assess 230 local communities’ vulnerability to disasters. Economic losses from disasters from 2001 to 2010 in Korea are analyzed using GIS. Moreover, this study examines the relationships between the constructed vulnerability indicators and economic damage from natural disasters. Machine learning techniques including Cubist and Random Forest are applied to examine what vulnerability indicators are statistically associated with disaster damage in Korea.