看语音情感识别看到的论文，论文是：
Survey on speech emotion recognition: Features, classification schemes,
and databases
其中有一段写到了HMM用作分类器的，原文中部分如下（版权问题我就不放原文全文了）：
The HMM classifier has been extensively used in speech
applications such as isolated word recognition and speech
segmentation because it is physically related to the production
mechanism of speech signal[102]. The HMM is a doubly stochastic
process which consists of a first-order Markov chain whose states
are hiddenfrom the observer. Associated with each state is a
random process which generates the observation sequence. Thus,
the hidden states of the model capture the temporal structure of
the data. Mathematically, for modeling a sequence of observable
data vectors,x1,…,xT, by an HMM, we assume the existence of a
hidden Markov chain responsible for generating this observable
data sequence. LetKbe the number of states,pi
, i¼1,y,Kbe the
initial state probabilities for the hidden Markov chain, and aij
,
i¼1,y,K,j¼1,y,Kbe the transition probability from stateito state
j. Usually, the HMM parameters are estimated based on the ML
principle. Assuming the true state sequence is s1,…,sT, the
likelihood of the observable data is given by
where
bi
ðxtÞPðxjst ¼iÞ
is the observation density of theith state. This density can be either
discrete for discrete HMM or a mixture of Gaussian densities for
continuous HMM. Since the true state sequence is not typically
known, we have to sum over all possible state sequences to find the
likelihood of a given data sequence, i.e.
====================================================
下面是我自己翻译的
====================================================
Categories