Method of computation by a computer of the probability of different sequences of arrangements of the observed state of a variable modeled by a cover of the markov model

摘要

The present invention relates to such a method, the hidden markov model comprising a set of states hidden, said computer comprising a processor and a memory, in which the processor performs the following tasks: a) to record the said sequences in a tree data stored in said memory, all the sequences, the number of states observed are identical to the rank t being grouped together on a same branch of the shaft, b) to calculate, for each node of the shaft and for each hidden state of the hidden markov model, a probability of observing the observed state of the arrangement of the observed state of said variable recorded in this first node when the model is in the state of the hidden at time 1 of this model, c) to calculate, for each node of depth t between 2 and t - 1, for each of the observed state of the u lying between 1 and p in the arrangement of states observed recorded in the depth t and, for each hidden state of the model, a probability of observing the - sequence of states, including, at each row single seen in between 1 and t, the observed state of the order of a u in the arrangement of states observed, corresponding to the sequence, when the model is in the state of the hidden at the time t of this model, and d) to calculate, for each terminal node of the shaft, the probability of the sequence of arrangements of states observed of said variable.