Supervised learning classifies labeled data

In supervised learning, a dataset with answers is provided. The machine then tries to classify that information.

For example, when wanting to create an algorithm that can classify beer from wine, you provide the alcohol content, colour and whether it is beer or wine.

One algorithm creates random bots to predict whether a beverage is beer or wine. The bots that passed the test with the highest score get additional variations while the inferior bots get discarded. Those modified bots take the test again. This process gets repeated until an algorithm is created that can predict the beverage with a high probability.

This sort of machine learning is an evolutionary algorithm since it resembles extremely fast Natural Selection.