So far we talked about classification
in terms of just predicting, is this a positive sentence or a negative
sentence, is email spam or not spam, but in general,
you wanna go a little bit beyond that and ask about, what is the probability
that this is email spam? So how confident am I on the prediction? So if you're just looking at positive and
negative, I want to know how sure I
am about the prediction. So for example,
if you take a sentence like, the sushi and everything else were awesome. That's a definite plus. Definitely positive. However, the sushi was good,
the service was okay. It's probably a plus but I'm not so sure. It's not as definite. And so what a classifier will often do,
is not just output is positive or negative, but output how confident,
how sure it is. One way to do that is to
talk about probabilities. So you have to play the probability
of being a positive or negative sentence,
given the input sentence x. So the output label, what's the probability output label,
given the input sentence? So for example, for the top example there,
instead of saying that's definite +, we say the probability that
it's a + given x in is 0.99. Even x is only 0.55 because I'm
kind of uncertain about that. Predicting probabilities or level of
confidence is extremely important and as we'll see in the classification course,
it allows you to do many things. So for example when you know the
probability, you can make decisions like, what is a good decision boundary
that trades off false positives and false negatives, and
balance between the two. [MUSIC]