题目阐释:
viterbi算法实现。 用python实现viterbi的hidden state 和 表现层的转移
动态规划问题,归结到
相邻两个step之间存在 state转移概率,state2emibission转移概率。
计算后可以得到每个step的每个state max_probablity
由于step_n依赖于 step_n-1,跟 step_n-2无关,所以可以一直如此往复,得到最后的max_prob
整个问题抽象为,下一个step依赖于上一个step的所有state,所以只需要计算每一层step的所有state的prbo即可。
难点:
三层for循环,为了保留,计算每个step的state的概率,所以要 next_state 嵌套在 source_state之外。
states=['Rainy','Sunny']
observations=['walk','shop','clean']
observations=('walk','clean','walk')
emission_probability={'Rainy':{'walk':0.1,'shop':0.4,'clean':0.5},
'Sunny': {'walk': 0.6, 'shop': 0.3, 'clean': 0.1}
}
trans_probability={'Rainy':{'Rainy':0.7,'Sunny':0.3},
'Sunny':{'Rainy':0.4,'Sunny':0.6}
}
start_probability={'Rainy':0.6,'Sunny':0.4}
def vertibi(states,objservations,start_prob,trans_prob,emi_prob):
T={state:[start_prob[state],[state],start_prob[state]] for state in states}
for objservation in objservations:
U={}
for next_state in states:
total=0
argmax=None
valmax=0
for source_state in states:
prob,v_path,v_prob=T[source_state]
p=emi_prob[source_state][objservation]*trans_prob[source_state][next_state]
prob*=p
v_prob*=p
if v_prob>valmax:
valmax=v_prob
argmax=v_path+[next_state]
total+=prob
U[next_state]=[total,argmax,valmax]
T=U
total = 0
argmax = None
valmax = 0
for state in states:
prob, v_path, v_prob=T[state]
if v_prob>valmax:
argmax=v_path
total=prob
valmax=v_prob
return total,argmax,valmax
total,argmax,valmax=vertibi(states,observations,start_probability,trans_probability,emission_probability)
print(total)
print(argmax)
print(valmax)