Confusion regarding Hidden Markov Model and Conditional Random Fields

Question

I am a bit confused about Hidden Markov Models and Conditional Random Fields. I wanna know id they are supervised or un-supervised learning methods? Thanks

This might get more interesting answers in the maths or AI stacks. — Charlie Martin, Jan 28 '11 at 15:32
Pull down the menu at top left of this page. Choose the 'all sites' button. Et voila. — Charlie Martin, Jan 29 '11 at 09:09

score 3 · Accepted Answer · answered Feb 02 '11 at 21:35

3

Well as I read several papers, they are both supervised methods and they need a labeled training set to be trained on.

answered Feb 02 '11 at 21:35

Hossein

40,161
57
141
175

score 2 · Answer 2 · answered Jan 28 '11 at 15:33

2

Neither. They're models for the underlying representation of knowledge. What happens during training is that certain transitions, being reinforced, become higher probability.

answered Jan 28 '11 at 15:33

Charlie Martin

110,348
25
193
263

1

Well do I need a labeled training set? – Hossein Jan 28 '11 at 15:42
I have a [similar](http://stackoverflow.com/questions/12015009/hidden-markov-model-training-for-dynamic-gestures) problem. Did you find any help @Hossein ? – masad Aug 18 '12 at 04:50

Confusion regarding Hidden Markov Model and Conditional Random Fields

2 Answers2