Hello
I have unbalanced data (neg 202, pos 48). Why the results are better when I switch the labels, pos 202 and neg 48. The data is balanced by choosing twice samples from the bigger class.
Malik
@Iris what do you think?
Can you add more details?
e.g what do you mean with twice samples from the bigger class? did you do this in the node or before?
This is a super small dataset. You also need to keep in mind the randomness of the random forest learners.
Hi @Iris
@stelfrich
@Iris
What do you think the problem ?
This topic was automatically closed 182 days after the last reply. New replies are no longer allowed.