Classification of landslide hazard/ susceptibility using Gaussian processes

Jahr:
2009

Autoren/Hrsg:
Dominik Gallus, Michael Ruff

Publikationstyp:
Kongressbeitrag/Proceeding

Quelle:
submitted to: Map World Forum/ Geo Science Forum (GSF), Hyderabad, India, 2009

Abstract:
Statistical classification techniques making use of GIS technology have been shown to yield good results at the task of an assessment of landslide hazard/ susceptibility on regional scale. In this paper, a machine learning technique known as the Gaussian process classifier (GPC), shown to yield classification performance comparable or superior to competing methods, is applied to the task. From a machine learning point of view, the advantages of an application of the technique to the task are discussed, and results of an application in a study area in the Northern Calcareous Alps (Vorarlberg, Austria) are presented.

zur Übersicht