I was researching about the supervised algorithm called Random Forest, that made me begin to study about decision trees, and how to induce them from a set, in order to create several predictors. My question comes at this point when we consider functions such as Information Gain or Gini impurity.

In order to use as splitting functions, both functions are very similar almost for the binary case, the only differences is that the first function has a maximum value of 1 whereas the other one has a maximum of 0.5 for the binary case. I would like to know in which cases it's is better to use one instead the another. I would also really appreciate any intuitive explanation of this kind of conclusion.

  • $\begingroup$ Also on cstheory: cstheory.stackexchange.com/questions/36550/…. $\endgroup$ – Yuval Filmus Sep 9 '16 at 15:52
  • $\begingroup$ @YuvalFilmus, I decided to post this question in the two sites since this question is in the middle point between practical and theoretical however I am open to suggestions to erase one and keep the other in the correct site, sorry for the inconvenients. $\endgroup$ – neo33 Sep 9 '16 at 20:20
  • $\begingroup$ Please do not post the same question on multiple sites. Each community should have an honest shot at answering without anybody's time being wasted. $\endgroup$ – D.W. Sep 13 '16 at 19:01
  • $\begingroup$ @D.W., I see but following your criteria what would be the correct site for this question, i really appreciate the support. $\endgroup$ – neo33 Sep 15 '16 at 16:26
  • $\begingroup$ CSTheory is only for research-level questions in theoretical CS. If it's not research-level, or not in theoretical CS, ask here. If it's research-level and relates to theoretical CS, pick one of the two sites, but not both. See also both sites help centers, meta.stackexchange.com/q/129598/160917, meta.cs.stackexchange.com/a/518/755. And just to be clear, it's a community-set policy. $\endgroup$ – D.W. Sep 16 '16 at 1:15

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Browse other questions tagged or ask your own question.