Hi I have a dataset with a few variables but two variables are very important. One of them is say S1 and other S2. S1 is a numeric variable and S2 is a categorical variable. Consider two cases of S1-S2 pair like S1 = 70.2, S2 A and S1 = 70.21, S2 = B. Assume my dataset has 20 instances of the first case and 20 instances of the second case. I have generated a synthetic class and used randomForest for calculating proximities between the 40 instances described above. The mds plot using this proximity doesnot seperate the first 20 instances from the next 20 instances. Any idea of why it doesn't work? What settings can I change in randomForest to separate these instances? Is there a better way to visualize the results than mds plots? I am looking at growing larger trees. How can i control the levels in the tree. Is nodesize the only parameter in randomForest to control it ? [[alternative HTML version deleted]]