목록에서 특수 요소를 선택하고 조건부 확률을 계산하십시오.

Question

리스트에서 해당 유니 그램이 존재하는 경우 bigram의 확률을 계산해야합니다. 원하는 결과는 예를 들어 다음 목록에 표시됩니다. 'pretty girl', 'pretty', 'girl'이 모두 존재합니다. 확률은, P 목록의 값을 사용하여, 그러므로 (0.0017) % (0.003 * 0.002) = 5.999999999999987e-06

S = ['girl', 'pretty', 'pretty girl', 'our', 'our world', 'wide', 'word', 'yes', 'yike', 'yummy'] 

P = [0.003, 0.002, 0.0017, 0.003, 0.006, 0.004, 0.002, 0.012, 0.006, 0.003] 

I가 다음 코드를 갖는다. 나에게 결과를주지 않는 것 같아서 나는 그 가능성을 계산하기 위해 앞으로 나아갈 수 없다. 이 코드로 수행하려고 시도하는 것은 목록에서 bigram을 선택하고 해당 unigram을 찾는 것입니다. 그런 다음 P에 확률을 매칭 할 계획입니다.

In [60]: import re 
In [61]: M = [] 
In [62]: for i in range(len(S)): 
      s_split = S[i].split() 
      s_split_len = len(S[i].split()) 
      if s_split_len == 2: 
       m = [] 
       a = re.match(s_split[0], S[i]) 
       b = re.match(s_split[1], S[i]) 
       m.append(a) 
       m.append(b) 
       M.append(m) 
       print M 

[[<_sre.SRE_Match object at 0x10447b988>, None], [<_sre.SRE_Match object at 0x10447b8b8>, None], [<_sre.SRE_Match object at 0x10447b920>, None], [<_sre.SRE_Match object at 0x10447b9f0>, None], [<_sre.SRE_Match object at 0x10447bac0>, None], [<_sre.SRE_Match object at 0x10447bb90>, None], [<_sre.SRE_Match object at 0x10447bbf8>, None], [<_sre.SRE_Match object at 0x10447bc60>, None], [<_sre.SRE_Match object at 0x10447bcc8>, None], [<_sre.SRE_Match object at 0x10447bd30>, None], [<_sre.SRE_Match object at 0x10447bd98>, None], [<_sre.SRE_Match object at 0x10447be00>, None], [<_sre.SRE_Match object at 0x10447be68>, None], [<_sre.SRE_Match object at 0x10447bed0>, None], [<_sre.SRE_Match object at 0x10447bf38>, None], [<_sre.SRE_Match object at 0x1044a8030>, None], [<_sre.SRE_Match object at 0x1044a8098>, None], [<_sre.SRE_Match object at 0x1044a8100>, None], [<_sre.SRE_Match object at 0x1044a8168>, None], [<_sre.SRE_Match object at 0x1044a81d0>, None], [<_sre.SRE_Match object at 0x1044a8238>, None], [<_sre.SRE_Match object at 0x1044a82a0>, None], [<_sre.SRE_Match object at 0x1044a8308>, None], [<_sre.SRE_Match object at 0x1044a8370>, None], [<_sre.SRE_Match object at 0x1044a83d8>, None], [<_sre.SRE_Match object at 0x1044a8440>, None], [<_sre.SRE_Match object at 0x1044a84a8>, None], [<_sre.SRE_Match object at 0x1044a8510>, None], [<_sre.SRE_Match object at 0x1044a8578>, None], [<_sre.SRE_Match object at 0x1044a85e0>, None], [<_sre.SRE_Match object at 0x1044a8648>, None], [<_sre.SRE_Match object at 0x1044a86b0>, None], [<_sre.SRE_Match object at 0x1044a8718>, None]] 
[[<_sre.SRE_Match object at 0x10447b988>, None], [<_sre.SRE_Match object at 0x10447b8b8>, None], [<_sre.SRE_Match object at 0x10447b920>, None], [<_sre.SRE_Match object at 0x10447b9f0>, None], [<_sre.SRE_Match object at 0x10447bac0>, None], [<_sre.SRE_Match object at 0x10447bb90>, None], [<_sre.SRE_Match object at 0x10447bbf8>, None], [<_sre.SRE_Match object at 0x10447bc60>, None], [<_sre.SRE_Match object at 0x10447bcc8>, None], [<_sre.SRE_Match object at 0x10447bd30>, None], [<_sre.SRE_Match object at 0x10447bd98>, None], [<_sre.SRE_Match object at 0x10447be00>, None], [<_sre.SRE_Match object at 0x10447be68>, None], [<_sre.SRE_Match object at 0x10447bed0>, None], [<_sre.SRE_Match object at 0x10447bf38>, None], [<_sre.SRE_Match object at 0x1044a8030>, None], [<_sre.SRE_Match object at 0x1044a8098>, None], [<_sre.SRE_Match object at 0x1044a8100>, None], [<_sre.SRE_Match object at 0x1044a8168>, None], [<_sre.SRE_Match object at 0x1044a81d0>, None], [<_sre.SRE_Match object at 0x1044a8238>, None], [<_sre.SRE_Match object at 0x1044a82a0>, None], [<_sre.SRE_Match object at 0x1044a8308>, None], [<_sre.SRE_Match object at 0x1044a8370>, None], [<_sre.SRE_Match object at 0x1044a83d8>, None], [<_sre.SRE_Match object at 0x1044a8440>, None], [<_sre.SRE_Match object at 0x1044a84a8>, None], [<_sre.SRE_Match object at 0x1044a8510>, None], [<_sre.SRE_Match object at 0x1044a8578>, None], [<_sre.SRE_Match object at 0x1044a85e0>, None], [<_sre.SRE_Match object at 0x1044a8648>, None], [<_sre.SRE_Match object at 0x1044a86b0>, None], [<_sre.SRE_Match object at 0x1044a8718>, None], [<_sre.SRE_Match object at 0x1044a8780>, None]] 

Answer 1

이

S = ['girl', 'pretty', 'pretty girl', 'our', 'our world', 'wide', 'world', 'yes', 'yike', 'yummy'] 

P = [0.003, 0.002, 0.0017, 0.003, 0.006, 0.004, 0.002, 0.012, 0.006, 0.003] 


for i in range(len(S)): 
    s_split = S[i].split() 
    s_split_len = len(S[i].split()) 
    if s_split_len == 2: 
     a = S.index(S[i]) 
     b = S.index(S[i].split()[0]) 
     c = S.index(S[i].split()[1]) 
     if a != None: 
      if b != None: 
       co = [a, b, c] 
       probs = P[co[0]], P[co[1]], P[co[2]] 
       print S[i], probs[0] % (probs[1] * probs[2])