Assume you’re dealing with an imbalanced dataset. I know I can do things like upsampling, downsampling, and synthetic sampling to build out my train and test split. My question is: if I’m using a random forest classifier, are there any implementations in R or Python that would force each of the randomly generated trees that it will be evaluating against such that it has balanced classes?
Solved – Imbalanced Classes: Random Forests w/ individually balanced trees
random forestunbalanced-classes
Best Answer
Apparently what this is describing is called "Balanced Random Forests."
There is a separate stack page that mentions a corresponding R package: Implementing Balanced Random Forest (BRF) in R using RandomForests