Skip to main content
. 2014 Nov 27;14(12):22500–22524. doi: 10.3390/s141222500

Figure 1.

Figure 1.

A visual comparison between a first order HMM and a linear chain CRF. The HMM defines a joint probability P(O, Q) whereas the CRF defines a conditional probability P(Q | O). Note that an HMM only has access to the current observation ot but the CRF has access to the entire observation sequence O at any given time.