| |
| |
Hidden Markov Models |
 | | A hidden Markov model (HMM) is a five-tuple (Omega_X,Omega_O,A,B,pi). |
 | | Suppose that we have a set W of words and a separate training set for each word. |
 | | The extension to HMMs with factored state spaces (e.g., see Figure 4) is graphically straightforward. |
| www.cs.brown.edu /research/ai/dynamics/tutorial/Documents/HiddenMarkovModels.html (1359 words) |
|