My name is Abhi and I am trying to teach myself data science by solving problems on the internet. In my current model I am using a random forest & the rfcv function to test the performance of the model.
My current understanding of this function is that this provides the cv error as we gradually reduce the number of features. But can some one help me answer the following questions
(a) How should I choose the value for the scale & step parameter? What factors influence this? (b) When I create the radomFores model there is a way to provide the nodeSize. How do I do this in the rfcv function
Thanks all. Any help is appreciated