Title

Authors

Abstract

The Federal Emergency Management Agency (FEMA) has spent more than 105.5 billion dollars in individual assistance just in 2018. Given the urban population growth in the country, especially the vulnerable coastal areas, this number tends to increase over time. Therefore, it is of crucial importance to optimize the spending of this budget. However, the mechanism of allocating individual assistance funds is mainly based on expert opinion and tacit knowledge, which can result in the insufficiency of the process. This research has studied the damage assessment report data on natural disasters in the United States published by FEMA between 2015 and 2020 to understand how much each state was affected by different types of natural disasters. The study uses socio-demographic variables for each state; the number of residents impacted, number of destroyed and damaged houses, type of the disaster, number of affected people, number of insured households, and number of low-income households as leading indicators to investigate the disaster resilience in each region. The study will also create supervised machine learning decision-making models based on Linear Regression, Decision Tree, and Support Vector Regression. These models will study the effect of the number of insured houses and low-income households on the number of individual assistant payments. The findings provide some insights that can help to understand how the funding is spent. The results can help the decision-makers, architecture, engineering, and construction teams gain insight into how they can make more resilient community.

 Keywords

Citation

Links