Aplicación de técnicas Big Data en la elaboración de mapas de peligrosidad de deslizamientos sismo-inducidos

Resum

We are currently immersed in a digital and industrial revolution, where data has acquired an undeniable importance. For this reason, geology has been positively affected by the current advances. One example of this is the implementation of new statistical learning techniques used to modelling and understanding complex datasets (Gareth et al., 2013), such as Big Data Analysis (BDA) or Machine Learning (ML) implemented in the field of geological hazards, in particular in the deve- lopment of seismic-induced landslide hazard maps. Some authors, such as Rodríguez-Peces et al. (2013), develop extensive seismic-induced landslide inventories, listing some characteristics, to obtain a set of factors that influence their occurrence. Thanks to these previous works, it has been shown that the binary classification of landslides making use of machine learning algorithms allow to classify the landslides that have occurred with greater accuracy. This classification constitutes one of the previous phases prior to obtaining landslide hazard maps, like the one made by Rodríguez-Peces et al. (2020), due to evaluate the precision of the model the results of the binary classification of landslides are used as a contrast element of a predicted hazard versus an occurred hazard.