Parameter tuning for Policy Learning
kayfoh opened this issue · comments
I would like to start off by saying good job on coming up with the Econml package. It has definitely been a useful tool to my data science toolkit. I have one question about the policy learning models that I was hoping to get some clarification on.
For the policy learning models available in econml (DRPolicyForest, DRPolicyTree, etc), I noticed that there is no score method available. My question is, how am I able to determine the best parameters eg, max_depth, n_estimators in the case of PolicyForest, if there is no score for me to evaluate against? Is there another way to determine the right parameter values?