Why is the complexity of Multiple Additive Regression Trees (MART) O(f * n_trees * d)
and not just O(n_trees * d)
? Doesn’t traversing a tree from top to bottom include all the necessary features?
have the same question
Why is the complexity of Multiple Additive Regression Trees (MART) O(f * n_trees * d)
and not just O(n_trees * d)
? Doesn’t traversing a tree from top to bottom include all the necessary features?
have the same question