2
$\begingroup$

I was learning about the Out of Bag error in random forests and I did not understand a point about the error calculation. Assume we have N bootstraps and there are a number of Out-Of-Bag samples for each one of these bootstraps.We can have a sample, say for instance sample x, appear in Out-Of-Bag part of multiple bootstraps. In such a case, do we consider these multiple x's as a single test sample or we will consider each one of them separately and say we have , for instance 5, test samples ( all being x), ? Thanks in advance.

Improve this question
$\endgroup$

1 Answer 1

Reset to default
1
$\begingroup$

I asked the question to the professor of the course I am taking. It turns out that when you have the Out-Of-Bag sample in multiple bootstraps, you only consider that sample once. For instance, the sample x occurred in Out-Of-Bag part of multiple bootstraps. When calculating the test set size , you consider these multiple x's as a unique (1) test sample.

Improve this answer
$\endgroup$

Your Answer

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

Not the answer you're looking for? Browse other questions tagged or ask your own question.