概率计算问题:已知λ\lambdaλ和O,求P(O∣λ)P(O|\lambda)P(O∣λ)根据概率公式:P(O∣λ)=∑IP(O,I∣λ)=∑IP(O∣I,λ)P(I∣λ) \begin{aligned} P(O|\lambda) & = \sum_{I}P(O, I | \lambda) \\ & = \sum_IP(O|I,\lambda)P(I|\lambda) \end{aligned} P(O∣λ)=I∑P(O,I∣λ)=I∑P(O∣I,λ)P(I∣λ) 当序列长度为T时:P(I∣λ)=πi1ai1i2⋯aiT−1iTP(O∣I,λ)=bi1(O1)⋯biT(OT) \begin{aligned} P(I|\lambda) = \pi_{i_1}a_{i_1i_2}\cdots a_{i_{T-1}i_T} \\ P(O|I,\lambda) = b_{i_1}(O_1)\cdots b_{i_T}(O_T) \end{aligned} P(I∣λ)=πi1ai1i2⋯aiT−1iTP(O∣I,λ)=bi1(O1)⋯biT(OT) I为所有可能的长度为T的状态序列的集合。
计算量大,不可行