Data–driven classification of landslide types at a national scale by using Artificial Neural Networks

• DOI: 10.1016/j.jag.2021.102549
• Publication Date: 2021-09-22T00:08:40Z
• Authors: Gabriele Amato; Lorenzo Palombi; Valentina Raimondi
• Keywords: Debris flow; Rockfall

Classification of landslide type is an essential step in risk management, although often missing large inventories.Here we propose a novel data-driven method that uses easily accessible morphometric and geospatial input parameters to classify landslides at national scale Italy by means shallow Artificial Neural Network.We achieved overall True Positive Rate 0.76 for five-class classification over 275,000 as (1) rockfall/toppling, (2) translational/rotational slide, (3) earth flow, (4) debris (5) complex landslide.In general, the model performance very good entire territory, with areas reaching F-score higher than 0.9.The can be applied any polygonal inventory, those produced automatic mapping procedures from Earth Observation imagery, order automatically identify types landslides.