It looks like you're new here. Sign in or register to get started.
What are the parameters of Gradient Boosted Trees that shoud be optimized, that can make a difference in performance?
Hi, technically any of the parameters can be optimized using the various optimization operators, but the parameters that have the largest impact in my experience are the number of trees, the maximal depth, the minimum split improvement, and the number of bins.
I would rather take learning rate than min_split