News for hackers. Stuff that matters.

A comment on Machine learning is easier than it looks. Parent comment

Re:Machine learning is easier than it looks by alextp on Thursday November 21, 2013 @09:29AM (#6774786)
Regarding the perceptron, most modern texts also have a regret analysis for the perceptron, which coupled with an online-to-batch conversion tells you how well do you expect a perceptron to perform on unseen data after a single pass, and it's usually a very good estimate (the answer is on average about as well as it did on the examples in the training data).
[ Parent | on HN ]
Re:Machine learning is easier than it looks by syllogism on Thursday November 21, 2013 @09:42AM (#6774866)
This isn't often discussed in my field (NLP) -- thanks!

We usually use averaged perceptron though, which seems like it would make this hard to analyse.

[ Parent | on HN ]
Re:Machine learning is easier than it looks by alextp on Saturday November 23, 2013 @10:26AM (#6786190)
The averaged perceptron is the one to which the proof applies :-)
[ Parent | on HN ]