Hi,
thank you for your work on this library.
Other implementations of random forests apply feature bagging at every split during tree generation. So I think it would be better to implement it at the CART level, not here.
Also, "replacement" is a bit of weird option, since it makes no sense to supply a feature more than once to the tree generation. Thus, it basically just randomly shrinks the number of features used for a tree.
Hi,
thank you for your work on this library.
Other implementations of random forests apply feature bagging at every split during tree generation. So I think it would be better to implement it at the CART level, not here.
Also, "replacement" is a bit of weird option, since it makes no sense to supply a feature more than once to the tree generation. Thus, it basically just randomly shrinks the number of features used for a tree.