The prediction in systems of pattern recognition based on hidden Markov models

The using of hidden markov models in variable subject-oriented systems of pattern recognition is represented in this paper. The algorithms of prediction based on these models are analyzed. The possible topologies of hidden markov models are considered. The example of recognition of sound patterns in restrict dictionary based on this model is discussed.