thanks, I`m trying it now. But it seems to work about 50 hours
The next problem is that RM sometimes failed to build a decision tree. I loop over examples (170 000 times) and after a few examples I design a decision tree build model. Within this loop it builds it cca 1600 times (that`s ok), but after 150th a decision tree failed and the model look like a one leaf and so every examples is labeled with one label, however there are two.
Is there a way to detect this behaviour simply? Can RM compare two models?
I have tried the group models, but I don`t know how it is worked. Does it take an average of confidence?