Application of Supervised Machine Learning Technique on LiDAR Data for Monitoring Coastal Land Evolution

Machine Learning (ML) techniques are now being used very successfully in predicting and supporting decisions multiple areas such as environmental issues land management. These have also provided promising results the field of natural hazard assessment risk mapping. The aim this work is to apply Supervised ML technique train a model able classify particular gravity-driven coastal hillslope geomorphic (slope-over-wall) involving most soft rocks Cilento (southern Italy). To model, only geometric data been used, namely morphometric feature maps computed on Digital Terrain Model (DTM) derived from Light Detection Ranging (LiDAR) data. Morphometric were using third-order polynomials, so obtain products that best describe landforms. Not all parameters literature significant ones chosen by applying Neighborhood Component Analysis (NCA) method. Different models trained main indicators confusion matrices compared. obtained Weighted k-NN (accuracy score = 75%). Receiver Operating Characteristic (ROC) curves shows discriminating capacity test reached percentages higher than 95%. resulting more accurate training area, will be extended similar along Tyrrhenian land.