What is the difference between supervised and unsupervised machine learning?
Comments
-
Labelled training data are necessary for supervised learning. For instance, you must first label the data you will need to train the model to classify the data into your labelled groups to do classification (a neglected learning job). Explicit data labelling is not necessary for unsupervised learning.
-
Supervised learning is like teaching a computer to recognize patterns by providing it with labeled examples. Just like training a dog with commands and rewards. For example, you teach the computer to recognize pictures of cats and dogs by showing it lots of cat and dog pictures labeled as "cat" or "dog." Then, when you show it a new picture, it can tell you if it's a cat or a dog based on what it learned.
Unsupervised learning, on the other hand, is more like letting the computer discover patterns on its own. You give it a bunch of things, but you don't tell it what they are. It's like giving a kid a pile of LEGO bricks without any instructions. The computer looks for patterns or groups by itself. For instance, if you give it a mix of pictures, it might group similar ones together without you saying why they're similar.
So, supervised learning is like giving the computer a clear path and teaching it specific things, while unsupervised learning is like letting the computer figure things out by itself, like a curious kid playing with toys. These both are the types of machine learning.
