spatial-ml-soc-2024

Code supplement: Spatial autocorrelation in machine learning for modelling soil organic carbon

Folders and files:

• data: reference to the input/source datasets and the conda environment yaml
• model_test: fresh training and (cross-)validation, numbers may slightly vary
• predict_full: scripts that were used to predict to the of Estonia as 10m raster

Errata:

In the manuscript the term RFSI (Random Forest Spatial Interpolation) is used, in the scripts and data the RFSI-associated model is named KNN (K-Nearest-Neighbours)