Information values of absolute elevation and elevation difference for illustration of thermal belt

This paper estimates the information gain and the information gain ratio, which are usually used in machine-learning processes, to assess which data layer – absolute elevation or elevation difference – better reflects the topoclimatic characteristics (especially the thermal belt). Both attributes are compared based on their information value in explaining the locations of vineyards, which depend largely on the thermal belt. The analysis is performed on 9,000 cells covering various winegrowing districts. In general, elevation difference proves to be a better attribute, but certain differences can be observed between individual areas, especially between the continental and submediterranean parts of Slovenia.