Expectation–maximization algorithm

The expectation-maximization algorithm is like a game of find and hide. It helps us find what we are looking for by pretending to hide it and then seeing how far it moves. First, the algorithm pretends to hide what we are looking for (called the parameters) in a complicated place. Then it looks at the data and sees if it can find the parameters. After this, the algorithm changes what it's hiding (the parameters) based on what it learned from looking at the data. It does this again and again until the parameters it is hiding are in the right place.