2
$\begingroup$

My goal is not to get better accuracy, but to make interpretation easier and have more stable splits.

So far most useful is use larger min_samples_leaf, but I always get nodes size almost same as min_samples_leaf.

Also I tried to cut input into interesting regions and run another tree in the subset.

Improve this question
$\endgroup$
1
  • 1
    $\begingroup$ If some variable is used only to split a few nodes out near the leaves, you might try eliminating it from your analysis. $\endgroup$
    – G5W
    May 12, 2019 at 17:15

0

Reset to default

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Browse other questions tagged or ask your own question.