Global ETD Search

1	Crisis Impact Prediction: A Data-driven Approach Paglamidis, Konstantinos January 2024 (has links) The field of crisis management and humanitarian assistance has been one of the major fields of development for governmental and common best European practices in the last decades. The European Union as a major humanitarian stakeholder has taken great effort to strengthen the response in case of humanitarian disasters. This work addresses the feasibility and possible benefits of using machine learning in the prediction of the impact severity of a disaster as a model-driven data analysis in comparison to data-driven reference models for early response coordination and preparedness. In comparison to classical data analysis systems the feasibility of earthquake impact prediction based on machine learning models is evaluated and further debated. machine learning impact prediction humanitarian disasters deep learning data mining artificial intelligence crisis management data analysis neural network decision tree random forest supervised learning CRISP-DM perceptron K-nearest neighbors hyper-parameter tuning natural disaster Computer and Information Sciences Data- och informationsvetenskap Computer Sciences Datavetenskap (datalogi)