Not-supervised training

The training not-supervised is a machine method of Learning. This method is distinguished from the Apprentissage supervised by the fact that there is no exit a priori . In the not-supervised training there is in entry a whole of collected data. Then the program treats these data like random variable and built a model of densities united for this unit of data.

The not-supervised training can also be used in conjunction with a Inférence bayésienne to produce conditional probabilities for each random variable being given the different ones.

Another form of not-supervised training is the Partitionnement of data which is not always probabilist.

See too

Random links:Corsier | Paroches | Ya `qûb bin Layth have-Saffâr | Slacken | Canadian museum of nature | Rick_Steiner