首先进行jieba分词，去除停用词；然后通过正则表达式去除无关字符，构建词向量；最后提取小说的所有人名并画图展示出来。

import jieba

import re

from gensim.models import Word2Vec

#读取数据

file = open(r'笑傲江湖.txt',encoding = 'utf-8')

text = file.readlines()

file.close()

#将换行符等特殊字符替换掉

text1 = text[1:] #第一行是这本小说的作者信息

text2 = [re.sub('\u3000| |\n','',i) for i in text1]

#分词,去除停用词

with open(r'停用词.txt','r',encoding = 'utf-8') as f:

stop_words = f.reanlines()

text_cut = [jieba.lcut(i) for i in text2]

stop_words = [re.sub(' |\n','',i) for i in stop_words]

text_ = [[i for i in word if i not in stop_words] for word in text_cut]

#构建词向量

my_wv = WordVec(text_,size = 200,min_counts = 5,window = 2,iter = 100)

#查看词向量

name = '令狐冲'

print(my_wv[name])

print(my_wv.similar_by_word(name,topn=3)) #查看跟'令狐冲'相关性前三的词

#查看跟令狐冲关系相当于师妹跟林平之的关系的词

my_wv.most_similar(['师妹','林平之'],[name],topn=3)

#查看跟令狐冲关系相当于师妹跟圣姑的关系的词

my_wv.most_similar(['师妹','圣姑'],[name],topn=3)

#查看跟令狐冲关系相当于师妹跟林平之的关系的词

import jieba.posseg as poss

#分词后对词的属性进行分析

a = poss.lcut('林平之是谁')

print(a)

print([list(i)[0] for i in a if list(i)[1] == 'nr'])

print(list(a[0])[1] == 'nr')

print(list(a[0])[0])

#提取这本小说里的所有人名

temp = [poss.lcut(i) for i in text2]

people = [[i.word for i in x if i.flag =='nr'] for x in temp]

temp2 = [' '.join(x) for x in people]

people2 = list(set(' '.join(temp2).split()))

data = []

newpeo = []

for i in people2:

try:

data.append(my_wv.wv[i])

newpeo.append(i)

except KeyError:

pass

#PCA降维后画图

from sklearn.decomposition import PCA

data2 = PCA(n_components=2).fit_transform(data)

print(len(data))

print(data2.shape)#(514，2)

import matplotlib.pyplot as plt

plt.rcParams['font.sans-serif'] = 'SimHei'

plt.rcParams['axes.unicode_minus'] = False

plt.figure(figsize=(15,15),dpi=360)

for i in range(len(data2)):

plt.scatter(data2[i,0],data2[i,1])

plt.text(data2[i,0],data2[i,1],newpeo[i])

plt.savefig('foo.png') #保存到本地

plt.show()

如果plt.show()显示的是空白的图的话就先将图片保存到本地，然后在打开图片查看即可。

中间那一堆密密麻麻的应该是跑龙套的演员，放大一部分查看。