0
$\begingroup$

I understand how pruning works in cross validation. But when the final tree is built, we will be using the entire dataset for tree training. How is the pruning performed then. I mean how do we see whether removing a leaf changes accuracy when there is no test dataset

$\endgroup$

1 Answer 1

1
$\begingroup$

Pruning is performed in the same way, when using the full data set. Unless you have not split the data set up into e.g. a training set and a test set (or use Cross Validation), you cannot test the performance of the tree in a proper way. I would thus highly recommend that you set some data aside to test on.

/Chris

$\endgroup$

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.