By this algorithm, we can evaluate the probability that a given HMM model produced a given observation sequence. Given an HMM model M , consisting of , , and , we compute the probability of the input sequence .
First, we define as the probability of generating the partial sequence , ending up in state j at time t . is initialized to 1.0 in the initial state, and 0.0 in all other states. If we have already computed for all i in the previous time frame t-1, then can be computed recursively in terms of the incremental probability of entering state j from each i while generating the output symbol (see Figure 2.5):
図 2.5: Forward pass recursion in HMM
If F is the final state, then we see that is the probability that the HMM generated the complete output sequence .