For multi-class classification problems, we use the cross entropy loss, which can be derived from a multinomial distribution via the maximum likelihoos estimation method.

I've already tried to derive this, but not succeeded yet. Any reference showing this would be appreciated!

  • $\begingroup$ Show us what you have tried so far $\endgroup$
    – nbro
    Jun 20 at 21:27


You must log in to answer this question.

Browse other questions tagged .