Viterbi Algorithms for Hidden Markov Models with Partially Visible States
In the paper,Viterbi algorithms for hidden Markov models are studied.When partial states,initial probability distributions,transition probability matrices and observation probability matrices are given,the optimal state sequences are esti-mated by the Viterbi algorithms.Compared with existing algorithms,the algorithms presented in the paper have not only considered the influence of partially visible states on the initial conditions and recursion formulas,but also ensured that the predicted state sequences are overall optimal.Finally,fault recognition is investigated to verify the feasibility of the algorithms.
NETL
NSTL
万方数据
