Vanneschi, LeonardoSoares, Rita da Silva2024-10-312024-10-312024-10-29http://hdl.handle.net/10362/174391Dissertation presented as the partial requirement for obtaining a Master's degree in Data Science and Advanced Analytics, specialization in Data ScienceIn this work, the potential of artificial intelligence for forecasting earthquakes in Portugal, a region under looked with only the Western Azores area being previously studied, is investigated. The target of this research is set to prediction of an event occurrence with a magnitude higher than 4.0 in the next 10 days. Seismic sequential data retrieved from an open-source catalogue is leveraged through machine learning models (Naïve Bayes, K-Nearest Neighbours, Support Vector Machine, Random Forest and neural networks focusing on the application of LSTM layers) to predict earthquake occurrences. The study involves an extensive data pre-processing with feature extraction by exploiting geophysical formulas and multiple strategiesto capture the sequential information. A study on the performance of each strategy and model is conducted in an attempt to identify if a methodology or algorithm should be preferred on earthquake studies in the area under analysis. The final results showcase some advancements in terms of state of the art and introduce new discoveries but still lack the minimum quality to consider the deployment of such warning systems on reallife. This thesis contributes to the development of AI-driven earthquake prediction systems in Portugal and suggests directions for future improvementsengEarthquake PredictionLSTMMachine LearningPortugalSequential DataSDG 3 - Good health and well-beingSDG 8 - Decent work and economic growthSDG 9 - Industry, innovation and infrastructureSDG 11 - Sustainable cities and communitiesmaster thesis203776763