Hi everybody, I'm using RF for ecological modelling. In particular I'm interested in defining crop suitable area. At the moment I have calibrated RF using a number of pedological and climatological data as predictor of presence/absence of crops. When plotting partial dependence of presence/absence versus a predictor, in some cases I obtained the graph as you see in file section (T_AUG.JPG) that represents the partial dependance on august mean temperature. It is not clear to me why the left side of the plot is flat and relatively high. Consider that in the dataset I used the presence of the crop has a minimum at 21?C. In the partial plot I would expect minimum values for temperature below 21?C. Could you help me? Thanking in advance Marco & Bob