[MUSIC] What we've covered so far is mostly a review of what you saw in
the first course of the specialization. At this point,
we're going to go much deeper into fundamental concepts which
are associated in logistical regression. In particular, in logistic regression, we
don't just predict plus one or minus one. We predict a probability. How likely is this review to be positive? How likely is it to be negative? And those probabilities are extremely
useful because they give us an indication of how sure we
are about predictions we make. So, in particular, so far we talked about taking the score,
throwing it into the sine function. Deciding is it a positive or
a negative review, so plus 1 or minus1. However, not all reviews
are created the same. So, for example, Rita says,
the sushi & everything was awesome! That's a definite plus 1,
I feel really good about that. But if Rita ays, the sushi was good, the service was okay,
I don't know what people think about that. It could be plus 1, it could be minus 1. If the person is really nice they might
feel bad about writing negative reviews, they might write that when they
really mean it's a negative review. So how can I capture the fact that for
the first one I'm sure that Y hat is plus one while for
the one on the right I might say. Y hat I don't know if its plus one or
minus one so I'm going to say Y hat is plus one probability 0.5 so
its either way its fine. [MUSIC]