matlab中hmm的使用示例

刚开始用matlab中的hmm函数的童鞋想必都对其参数或者返回值或多或少会有一些疑问，matlab官方文档 http://www.mathworks.cn/cn/help/stats/hidden-markov-models-hmm.html 对此的解释又是相当简洁，本文写了一些博主对这些函数的理解，如果有什么疑问或者指正的话欢迎提出O(∩_∩)O~

1. hmmdecode

1.1 matlab的运行

>> t = [.7, .3; .4, .6]
t =   %transmission  matrix
    0.7000    0.3000
    0.4000    0.6000

>> e = [.1, .1 .8; .8, .1, .1]
e =  %emission matrix
    0.1000    0.1000    0.8000
    0.8000    0.1000    0.1000

>> s = [1, 2, 3]  %sequence
s =
     1     2     3

>> [p,logPseq,fs,bs,scale] = hmmdecode(s,t,e)
p =  % calculates the posterior state probabilities  of the sequence  seq
    0.2488    0.5771    0.9039
    0.7512    0.4229    0.0961

logPseq =  % the logarithm of the probability of sequence  seq
   -4.2114

fs =  %fs and bs returns  forward and backward probabilities of the sequence scaled by  S .
    1.0000    0.2258    0.4677    0.9039
         0    0.7742    0.5323    0.0961

bs =
    1.0000    1.1020    1.2337    1.0000
    1.6445    0.9703    0.7946    1.0000

scale =
    1.0000    0.3100    0.1000    0.4782

1.2 返回值验证

写程序对返回值进行验证，程序见文后附录。