Citation

Fernandez R, Correal JF, D'Ayala D, Medaglia AL. Int. J. Disaster Risk Reduct. 2023; 90: e103655.

Copyright

(Copyright © 2023, Elsevier Publishing)

DOI

10.1016/j.ijdrr.2023.103655

PMID

unavailable

Abstract

Education infrastructure is one of the main barriers on school quality in Low- and Middle-Income Countries (L&MICs), since it is insufficient and unevenly distributed. Improving the school infrastructure is needed to provide a high-quality education environment. Although research on how to improve the infrastructure is available, there is still a lack of a consistent and systematic approach to develop large-scale interventions at the national or regional level. To fill this gap, we propose a data-driven methodology with the purpose of developing a prioritization of interventions to carry out a seismic disaster risk reduction program. The method starts by identifying groups of similar buildings using clustering analysis, starting with a seismic taxonomy as descriptor (i.e., model input). Then, domain experts analyze the suggested clusters to design scalable interventions for the representative building of each cluster. The proposed data-driven methodology requires experts' criteria in each step to validate the results and make them applicable, but significantly reduces the bias by automating the decision-making process. We use as case study the Dominican Republic public school infrastructure and present the results of the application of the proposed method. The method presented herein is extensible to other infrastructure portfolios, as well as to other types of hazards.

Language: en

Keywords

Clustering analysis; Disaster mitigation; Disaster preparedness; Education; Machine learning; School infrastructure; Unsupervised machine learning