How can we get the "accuracy" of a classification

Hi guys, in the video Supervised learning part 2 at 4:19 Andrew said that classification can’t output numbers between 0 and 1 (there is only two category).

So why do I see a lot of classification examples that output the percentage of the possible object in an image ? They do need more than just 0 or 1 right ? Or maybe they just don’t use unsupervised learning for this type of classification.

Thank you!

1 Like

The classifications are based on finding the output that has the highest probability.

In the examples you see where the probability is shown, those are one step before the classification is predicted.

1 Like