Methods and Challenges in Digital Soil Mapping: Applied Modelling with R Examples
摘要
Advances in software and hardware have made machine-learning-based digital soil mapping (DSM) a new and extensively researched solution in soil surveying. However, the high complexity of soil systems and their subsurface nature are still major challenges to DSM. To spread recently developed approaches to DSM, we systematically described the steps to be followed in a DSM project using machine learning techniques. We also presented practical and open data examples using R language and discussed some aspects related to machine learning modelling in the natural sciences.