Machine learning is all the rage. What does /g/ know about

>>57367011
>What does /g/ know about machine learning?

I know that it's all the rage.

I hear it's all the rage.

File: trump bird.jpg (284KB, 865x1200px) Image search: [Google]

284KB, 865x1200px

>>57367023
>>57367036
you're fired

>>57367011
nice popper

>>57367011
>What does /g/ know about machine learning?
Markov chains and shit

>>57367562
whats a popper

>>57367574
a dogger

>>57367585
oh thats nice

I'm not well versed, but a team mate of mine is. We're in the process of using machine learning to take performance metrics and other data from servers (we have many thousands) to see if we can predict which servers are candidates for decommissioning. We assume we have a lot of unused servers, especially in non-prod.

File: 08 - e2plAZf.gif (44KB, 257x293px) Image search: [Google]

44KB, 257x293px

>>57367986
neat

>>57367011
Machine learning is pretty much nothing but applied probability.

You have a huge dataset.
You determine the features.
You create a bunch of training instances.
You run the machine learning algorithm on the training instances which will then create a model based on the probabilities.
You then use the model to predict new instances.