What I know about CRF is that they are discriminative models, while HMM are generative models, but, in the inference method, both use the same algorithm, that is, the Viterbi algorithm, and forward and backward algorithms.

Does CRF use the same features as HMM, namely features transition and state features?

But in here https://homepages.inf.ed.ac.uk/csutton/publications/crftut-fnt.pdf, CRF has these features Edge-Observation and Node-Observation Features.

What is the difference features transition and state features vs features Edge-Observation and Node-Observation features?

  • $\begingroup$ crf is undirected model and hmm is a type of bayesian net which is directed (+ has additional constraints making computation over it easier) $\endgroup$ – mshlis Jul 31 '19 at 14:55

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Browse other questions tagged or ask your own question.