3 ms·
Well, in that case minimizing the (negative) loglikelihood seems principled but you could minimize literally any loss function and it would work almost just as
by kgwgk 7y ago
Well, in that case minimizing the (negative) loglikelihood seems principled but you could minimize literally any loss function and it would work almost just as well.
- mlevental 7y agoLol agreed!