j
k
j a
j l
*
There is no problem, in general, with overfitting, as long as your evaluation of an estimator's performance isn't biased towards the training set. We've not talked about evaluation.
Attachments:
Back to the thread
Back to the list
