Large-Scale Machine Learning at Twitter [pdf]

Re:Large-Scale Machine Learning at Twitter [pdf] by ot on Tuesday June 12, 2012 @07:46PM (#4103542)

I don't see why it should have equivalent modeling power, the set of decision surfaces that it can represent is strictly larger.

[ Parent | on HN ]

Re:Large-Scale Machine Learning at Twitter [pdf] by bravura on Tuesday June 12, 2012 @08:43PM (#4103723)

puts on theoretician's hat

You are correct, it can represent more decision surfaces.

takes off theoretician's hat

In practice, my intuition tells me that aggregating over an ensemble of logreg models trained on subsets of the data is inferior to training a single logreg model over the entire data.

I apologize for being sloppy tho.

[ Parent | on HN ]