Machine learning and processing of large data

Machine learning refers to a set of tools to establish models of linear or non-linear relations or other previously unknown relationships in complex data. In soil science it is widely used to create soil maps, to derive difficult to measure soil properties from sensor data or to provide information for soil related decision-making in agriculture. Machine learning algorithms based on decision trees are among the most popular techniques. Models with good predictive power are complex, but often outperform simpler models from classical statistics. Interpretation with algorithms that subsequently analyze the models is possible yet challenging.