6.1 文本表示（one-hot）

one-hot原理：https://blog.csdn.net/liusisi_/article/details/107149750

代码：

import jieba
import numpy as np
from collections import Counter

stopwords = [line.strip() for line in open('../data/stopwords.data', 'r', encoding='utf-8').readlines()]

sentence1 = '小明喜欢看电影小红喜欢听歌'
sentence2 = '小张喜欢打篮球'
text = sentence1 + sentence2

fenci_result = jieba.cut(text)
# for word in stopwords:
# 	if word in fenci_result:
# 		fenci_result.remove(word)
dict_result = Counter(fenci_result)
dict_text = {}
index = 0
for dic in dict_result:
	dict_text[dic] = index
	index += 1
print(dict_text)
fenci_sentence1 = jieba.cut(sentence1)
fenci_sentence2 = jieba.cut(sentence2)
vec1 = [0] * len(dict_text)
vec2 = [0] * len(dict_text)
for dic in dict_text:
	for s in fenci_sentence1:
		index = dict_text[s]
		vec1[index] += 1
	for s in fenci_sentence2:
		index = dict_text[s]
		vec2[index] += 1
print(vec1)
print(vec2)

运行结果：

相似度计算：

#余弦相似度
sum_fz = 0
distance_cos = 0
sum_vec1 = 0
sum_vec2 = 0
sum_fm = 0
for i in range(len(vec1)):
	sum_fz += vec1[i] * vec2[i]
	sum_fm += math.sqrt(math.pow(vec1[i], 2)) + math.sqrt(math.pow(vec2[i], 2))

distance_cos = sum_fz / sum_fm
print('余弦值：', distance_cos)  # 余弦值
print('角度值', math.acos(distance_cos))  # 角度值