EM algorithm: how it works

Full lecture: http://bit.ly/EM-alg
Mixture models are a probabilistically-sound way to do soft clustering. We assume our data is sampled from K different sources (probability distributions). The expectation maximisation (EM) algorithm allows us to discover the parameters of these distributions, and figure out which point comes from each source at the same time.

Видео EM algorithm: how it works канала Victor Lavrenko