Hidden Markov Models allow for training of lambda = (A, B, pi) simply by supplying the observations and number of hidden states. To allow for this prediction on unknown state should be a central task of the model
It was decided, that in order to meet the contrains given by the scope of the thesis to omit the ability of the tool to predict a truly "hidden" state.
Hidden Markov Models allow for training of lambda = (A, B, pi) simply by supplying the observations and number of hidden states. To allow for this prediction on unknown state should be a central task of the model