18、不确定推理中的隐马尔可夫模型与动态信念网络

不确定推理中的隐马尔可夫模型与动态信念网络

1. 隐马尔可夫模型（HMM）

1.1 精确滤波

在HMM中， HMMVEfilter 类用于精确滤波。它维护一个当前状态分布，可通过观察或推进到下一个时间步来更新。以下是该类的代码实现：

from display import Displayable

class HMMVEfilter(Displayable):
    def __init__(self,hmm):
        self.hmm = hmm
        self.state_dist = hmm.indist

    def filter(self, obsseq):
        """updates and returns the state distribution following the
        sequence of
        observations in obsseq using variable elimination.

        Note that it first advances time.
        This is what is required if it is called sequentially.
        If that is not what is wanted initially, do an observe first.
        """
        for obs in obsseq:
            self.advance()  # advance tim