Machine learning approach for disaster risk and resilience assessment in coupled human infrastructure systems performance

There is a gap in the literature on data-driven analyses for post-disaster evaluation of community risk and resilience, particularly in utilizing features related to the performance of coupled human-infrastructure systems. This study developed an index and machine learning-based method for assessing community risk and resilience after a disaster. Using feature groups related to population protective actions, infrastructure/building performance, and recovery features, the study examined risk and resilience performance in communities affected by Hurricane Harvey in Harris County, Texas, in 2017. It analyzed disparities across four archetypes of risk-resilience status, and income groups, revealing how spatial areas are shaped by the performance of coupled human-infrastructure systems. The findings also highlight the complex relationship between socio-economic factors, risk exposure, and resilience. This study provides researchers and practitioners with new data-driven and machine intelligence-based methods to evaluate community risk and resilience during disasters, offering insights to inform future policies and decision-making.