1. 先下载并安装nltk包，准备一张简单的图片存入代码所在文件目录，搜集英文停用词表

import nltk

nltk.download()

2. 绘制词云图

import re

import numpy as np

import pandas as pd

#import matplotlib

import matplotlib.pyplot as plt

from nltk.tokenize import word_tokenize, sent_tokenize

from nltk.corpus import stopwords

from PIL import Image

from wordcloud import WordCloud

from sklearn.datasets import fetch_20newsgroups

#from sklearn.feature_extraction.text import CountVectorizer

from collections import Counter, defaultdict

def word_cut(contents, cut=','):

res = []

for content in contents:

content = content.lower()

words = [word for word in re.split(cut, content) if word]

res.append(words)

return res

def word_count(contents):

#words_count = Counter(sum(contents,[])) #慢

word_count_dict = defaultdict(lambda: 0)

for content in contents:

temp_dict = Counter(content)

for key in temp_dict:

word_count_dict[key] += temp_dict[key]

return word_count_dict

def stopwords_filter(contents, stopwords):

contents_clean = []

word_count_dict = defaultdict(lambda: 0)

for line in contents:

line_clean = []

for word in line:

if word in stopwords:

continue

line_clean.append(word)

word_count_dict[word] += 1

contents_clean.append(line_clean)

words_count = list(word_count_dict.items())

words_count.sort(key=lambda x:-x[1])

words_count = pd.DataFrame(words_count, columns=['word', 'count'])

return contents_clean, words_count

从外部导入数据

'''

df_news = pd.read_table('val.txt', names=['category','theme','URL','content'], encoding='utf-8')

stopwords = pd.read_csv("stopwords.txt", index_col = False, sep="\t",

quoting=3, names=['stopword'], encoding='utf-8')

contents = df_news.content.values.tolist()

stopwords = stopwords.stopword.values.tolist()'''

自定义切词

'''

#[ ,.\n\t--':;?!/+<>@]

#[ ,.\n\t=--'`_:;?!^/|+<>{}@~\ ]

#contents = word_cut(contents=news.data, cut='[ ,.\n\t-`_:;?!^/|+<>{}@~]')

'''

将数据整理为模型入参形式

'''

#vec = CountVectorizer()

#X_train = vec.fit_transform(X_train) #不可直接将vec用在测试集上

#vectorizer_test = CountVectorizer(vocabulary=vec.vocabulary_)

#X_test = vectorizer_test.transform(X_test)

'''

可从中筛选停用词

'''

word_count_dict = word_count(contents)

temp = list(word_count_dict.items())

temp.sort(key=lambda x:-x[1])

df = pd.DataFrame(temp, columns=['word','count'])

df.to_csv(r'D:\PycharmProjects\zsyb\stop_words.csv')

'''

调包实现上述功能

news = fetch_20newsgroups(subset='all')

自定义的快好几倍，可以加if not in ‘’去标点

contents = [word_tokenize(content.lower()) for content in news.data] #sent_tokenize(content)

punctuations = set(list(',.\n\t-`_():;?!$#%&.*=^/|+<>{}@~')) #标点

digits = {str(i) for i in range(50)}

others = {'--', "''", '``', "'", '...'}

下载网上的停用词表加入 nltk_data\corpora\stopwords，低频词过滤（不要加入停用词）

stopWords = set(stopwords.words('english')) | punctuations | digits | others

contents_clean, words_count = stopwords_filter(contents, stopWords)

#df.groupby(by=['word']).agg({"count": np.size})

绘制词云图

fontpath = 'simhei.ttf'

aimask = np.array(Image.open(r"D:\PycharmProjects\zsyb\pig.png"))

wc = WordCloud(font_path = fontpath, #设置字体

background_color = "white", #背景颜色

max_words = 1000, #词云显示的最大词数

max_font_size = 100, #字体最大值

min_font_size = 10, #字体最小值

random_state = 42, #随机数

collocations = False, #避免重复单词

mask = aimask, #造型遮盖

width = 1200, height = 800, #图像宽高，需配合plt.figure(dpi=xx)放缩才有效

margin = 2 #字间距

)

word_frequence = {x[0]:x[1] for x in words_count.head(100).values}

word_cloud=wc.fit_words(word_frequence)

plt.figure(dpi=100) #通过这里可以放大或缩小

plt.subplot(121)

plt.imshow(aimask)

#plt.axis("off") #隐藏坐标

plt.subplot(122)

plt.imshow(word_cloud)

#plt.axis("off") #隐藏坐标