Closed ZihanNiu closed 7 years ago
We've assumed that the training and testing data are drawn from the same distribution. So, a classifier that does well on the training data should also do well on the testing data. (Though, this won't always be the case, as we'll discuss through the remainder of the course.)
On Feb 24, 2017, at 4:58 PM, ZihanNiu notifications@github.com wrote:
Hi Prof, I got a question on this concept, why we need to maximize the joint probability of the true labels for all training instances.
— You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub, or mute the thread.
Hi Prof, I got a question on this concept, why we need to maximize the joint probability of the true labels for all training instances.