10.18453/rosdok_id00004228 Hagenauer, Julian Christian Julian Christian Hagenauer http://d-nb.info/gnd/1036794342 Universität Rostock 2022 004 Data processing Computer sciences 550 Earth sciences 2022 en http://purl.uni-rostock.de/rosdok/id00004228 urn:nbn:de:gbv:28-rosdok_id00004228-3 The main objective of this thesis is to improve the usefulness of spatial machine learning for the spatial sciences and to allow its unused potential to be exploited. To achieve this objective, this thesis addresses several important but distinct challenges which spatial machine learning is facing. These are the modeling of spatial autocorrelation and spatial heterogeneity, the selection of an appropriate model for a given spatial problem, and the understanding of complex spatial machine learning models.