Autor: RENATO JOSE QUILICHE ALTAMIRANO
Orientadores: Adriana Leiras – PUC-Rio & Fernanda Araújo Baião Amorim – PUC-Rio
Data e Hora: 04/09/2023, 14h
Link/ Sala: https://puc-rio.zoom.us/j/91470014337?pwd=ak1iM3pSVzBRNHRtZEJRTnk2c0tqQT09
ID da reunião: 914 7001 4337
Senha de acesso: 803153
Banca Examinadora: Adriana Leiras – orientador – PUC-Rio; Fernanda Araújo Baião Amorim – coorientadora – PUC-Rio; Mario Chong – Universidad del Pacífico; Paula Medina Maçaira Louro – PUC-Rio; Irineu de Brito Junior – UNESP .
Resumo:
This dissertation presents a data-driven approach to the problem of recurrent disasters in developing countries. Supervised machine learning methods are used to train classifiers that aim to predict whether a family would be affected by recurrent weather threats (one classifier is trained for each natural hazard). The developed approach is valid for recurrent natural hazards that affect a country and
allows disaster risk managers to accurately target their operations. In addition, predictive evaluation allows managers to understand the drivers of these predictions, leading to proactive policy formulation and operations planning to mitigate risks and prepare communities for recurrent disasters. The proposed methodology was applied to the case study of Peru, where classifiers were trained for Cold waves, Floods, and Landslides. In the case of Cold waves, the classifier has 73.82% accuracy. In the case of Floods, 82.57% of accuracy was achieved. In the case of Landslides, 88.85% accuracy was achieved. The classifiers provide an intelligent data-driven method that saves resources while ensuring accuracy. Prediction is driven by
households’ wealth and geographical location. In addition, the research provides guidelines for practical implementations of the classifiers, including the use in planning of humanitarian supply network design.
The research results have several managerial implications, so the authors call on disaster risk managers and other relevant stakeholders to take action. Recurrent disasters challenge all humanity.