自然語言處理之比較兩個句子的相似度 余弦相似度

1.句子如下：

s1 = "周傑倫是一個歌手,也是一個叉叉"
s2 = "周傑倫不是一個叉叉，但是是一個歌手"

2.分詞：

import jieba
s1_list = [x for x in jieba.cut(s1,cut_all=True) if x != '']
s2_list = [x for x in jieba.cut(s2,cut_all=True) if x != '']

print (s1_list)
print (s2_list)

2.詞頻向量化：

import jieba
s1 = "周傑倫是一個歌手,也是一個叉叉"
s2 = "周傑倫不是一個叉叉，但是是一個歌手"

s1_list = [x for x in jieba.cut(s1,cut_all=True) if x != '']
s2_list = [x for x in jieba.cut(s2,cut_all=True) if x != '']
s1_set = set(s1_list)
s2_set = set (s2_list)
word_dict=dict()
i=0
for word in  s1_set.union(s2_set):
   word_dict [word] = i
   i+=1

#詞頻向量化函數，輸入一個總字典和待向量化列表
def word_to_vec(word_dict,lists):
    word_count=dict()
    result=[0]*len(word_dict)
    for word in lists:
        if word_count.get(word,-1) == -1:
            word_count[word]=1
        else:
            word_count[word]+=1
    for word,freq in word_count.items():
        wid = word_dict[word]
        result[wid]=freq
    return result
s1_vec=(word_dict,s1_list)
s2_vec=(word_dict,s2_list)
print (s1_vec)
print (s2_vec)

4.計算2個向量的相似度：

import jieba
import math
s1 = "周傑倫是一個歌手,也是一個叉叉"
s2 = "周傑倫不是一個叉叉，但是是一個歌手"

s1_list = [x for x in jieba.cut(s1,cut_all=True) if x != '']
s2_list = [x for x in jieba.cut(s2,cut_all=True) if x != '']
s1_set = set(s1_list)
s2_set = set (s2_list)
word_dict=dict()
i=0
for word in  s1_set.union(s2_set):
   word_dict [word] = i
   i+=1
#詞頻向量化函數
def word_to_vec(word_dict,lists):
    word_count=dict()
    result=[0]*len(word_dict)
    for word in lists:
        if word_count.get(word,-1) == -1:
            word_count[word]=1
        else:
            word_count[word]+=1
    for word,freq in word_count.items():
        wid = word_dict[word]
        result[wid]=freq
    return result
#計算2個向量的余弦相似度
def cos_dist(a, b):
    if len(a) != len(b):
        return None
    part_up = 0.0
    a_sq = 0.0
    b_sq = 0.0
    for a1, b1 in zip(a,b):
        part_up += a1*b1
        a_sq += a1**2
        b_sq += b1**2
    part_down = math.sqrt(a_sq*b_sq)
    if part_down == 0.0:
        return None
    else:
        return part_up / part_down
s1_vec = word_to_vec(word_dict,s1_list)
s2_vec = word_to_vec(word_dict,s2_list)
num=cos_dist(s1_vec,s2_vec)
print (num)