关键词提取使用的是人民网的粤经济新闻数据,分别实现基于TF-IDF、TextRank和Word2vec词聚类的关键词提取算法。该数据集共包含558个文本文件,每个文件的内容均为标题和摘要。
关键词提取的实现流程
(1)将原始数据处理成result.csv文本,具体编号、标题和摘要,Text.csv数据格式如图所示。
(2)获取每行记录的标题和摘要字段,并拼接这两个字段。
(3)加载自定义停用词表stopWord.txt,然后对拼接的文本进行数据预处理操作,包括分词、去除停用词、用空格分隔文本等。
(4)编写相应算法提取关键词。
(5)将最终结果写进文件进行保存。
项目地址:zz-zik/NLP-Application-and-Practice: 本项目将《自然语言处理与应用实战》原书中代码进行了实现,并在此基础上进行了改进。原书作者:韩少云、裴广战、吴飞等。 (github.com)
数据预处理
基于TF-IDF的关键词提取算法
data_prepare.py:数据预处理,合并文本文件,在data文件夹下生成text.csv文件,内容包括,id、title、key
import os
import csv
# 文本文件合并
def text_combine(path):
# 1.获取文件列表
files = []
for file in os.listdir(path):
if file.endswith(".txt"):
files.append(path+"/"+file)
# 2.创建text.csv文件,保存结果
with open('data/text.csv','w',newline = '',encoding= 'utf-8') as csvfile:
writer = csv.writer(csvfile)
writer.writerow(['id','title','abstract'])
# 3.遍历txt文件,获取文件编号
for file_name in files:
number = (file_name.split('/')[1].split('_')[0])
title,text = '',''
count = 0
# 4.读取标题和内容
with open(file_name,encoding='utf-8-sig') as f:
for line in f:
if count == 0:
title += line.strip()
else:
text += line.strip()
count += 1
res = [number,title,text]
writer.writerow(res)
# 主函数处理
def main():
path = 'text_file'
text_combine(path)
if __name__ == '__main__':
main()
基于TextRank的关键词提取算法
tfidf.py实现基于TextRank的关键词提取算法
# coding=utf-8
import codecs
import pandas as pd
import numpy as np
# 导入jieba分词
import jieba.posseg
import jieba.analyse
# 导入文本向量化函数
from sklearn.feature_extraction.text import TfidfTransformer
# 导入词频统计函数
from sklearn.feature_extraction.text import CountVectorizer
# 读取text.csv文件:分词,去停用词,词性筛选
def data_read(text, stopkey):
l = []
pos = ['n', 'nz', 'v', 'vd', 'vn', 'l', 'a', 'd'] # 定义选取的词性
seg = jieba.posseg.cut(text) # 分词
for i in seg:
if i.word not in stopkey and i.flag in pos: # 去停用词 + 词性筛选
l.append(i.word)
return l
# tf-idf获取文本top10关键词
def words_tfidf(data, stopkey, topK):
idList, titleList, abstractList = \
data['id'], data['title'], data['abstract']
corpus = [] # 将所有文档输出到一个list中,一行就是一个文档
for index in range(len(idList)):
# 拼接标题和摘要
text = '%s。%s' % (titleList[index], abstractList[index])
text = data_read(text, stopkey) # 文本预处理
text = " ".join(text) # 连接成字符串,空格分隔
corpus.append(text)
# 1、构建词频矩阵,将文本中的词语转换成词频矩阵
vectorizer = CountVectorizer()
# 词频矩阵,a[i][j]:表示j词在第i个文本中的词频
X = vectorizer.fit_transform(corpus)
# 2、统计每个词的tf-idf权值
transformer = TfidfTransformer()
tfidf = transformer.fit_transform(X)
# 3、获取词袋模型中的关键词
word = vectorizer.get_feature_names()
# 4、获取tf-idf矩阵,a[i][j]表示j词在i篇文本中的tf-idf权重
weight = tfidf.toarray()
# 5、打印词语权重
ids, titles, keys = [], [], []
for i in range(len(weight)):
print(u"-------这里输出第", i + 1, u"篇文本的词语tf-idf------")
ids.append(idList[i])
titles.append(titleList[i])
df_word, df_weight = [], [] # 当前文章的所有词汇列表、词汇对应权重列表
for j in range(len(word)):
print(word[j], weight[i][j])
df_word.append(word[j])
df_weight.append(weight[i][j])
df_word = pd.DataFrame(df_word, columns=['word'])
df_weight = pd.DataFrame(df_weight, columns=['weight'])
word_weight = pd.concat([df_word, df_weight], axis=1) # 拼接词汇列表和权重列表
word_weight = word_weight.sort_values(by="weight", ascending=False) # 按照权重值降序排列
keyword = np.array(word_weight['word']) # 选择词汇列并转成数组格式
word_split = [keyword[x] for x in range(0, topK)] # 抽取前topK个词汇作为关键词
word_split = " ".join(word_split)
keys.append(word_split.encode("utf-8").decode("utf-8"))
result = pd.DataFrame({"id": ids, "title": titles, "key": keys},
columns=['id', 'title', 'key'])
return result
def main():
# 读取数据集
dataFile = 'data/text.csv'
data = pd.read_csv(dataFile)
# 停用词表
stopkey = [w.strip() for w in codecs.open('data/stopWord.txt', 'r', encoding="utf-8").readlines()]
# tf-idf关键词抽取
result = words_tfidf(data, stopkey, 10)
result.to_csv("result/tfidf.csv", index=False)
if __name__ == '__main__':
main()
构建文本数据的词向量
textrank.py:根据wiki.zh.text.vector词向量模型构建文本数据的词向量,并获取候选词语的词向量。
# coding=utf-8
import pandas as pd
import jieba.analyse
# 处理标题和摘要,提取关键词
def words_textrank(data, topK):
idList, titleList, abstractList = data['id'], data['title'], data['abstract']
ids, titles, keys = [], [], []
for index in range(len(idList)):
# 拼接标题和摘要
text = '%s。%s' % (titleList[index], abstractList[index])
jieba.analyse.set_stop_words("data/stopWord.txt") # 加载自定义停用词表
print("\"", titleList[index], "\"", " 10 Keywords - TextRank :")
# TextRank关键词提取,词性筛选
keywords = jieba.analyse.textrank(text, topK=topK,
allowPOS=('n', 'nz', 'v','vd', 'vn','l', 'a', 'd'))
word_split = " ".join(keywords)
keys.append(word_split.encode("utf-8").decode("utf-8"))
ids.append(idList[index])
titles.append(titleList[index])
result = pd.DataFrame({"id": ids, "title": titles,
"key": keys},
columns=['id', 'title', 'key'])
return result
def main():
dataFile = 'data/text.csv'
data = pd.read_csv(dataFile)
result = words_textrank(data, 10)
result.to_csv("result/textrank.csv", index=False)
if __name__ == '__main__':
main()
基于Word2vec词聚类的关键词提取算法
实现基于Word2vec词聚类的关键词提取算法
(1)编写word2vec_prepare.py,构建候选词向量
# coding=utf-8
import warnings
warnings.filterwarnings(action='ignore',
category=UserWarning,
module='gensim') # 忽略警告
import codecs
import pandas as pd
import numpy as np
import jieba # 分词
import jieba.posseg
import gensim # 加载词向量模型
# 返回特征词向量bai
def word_vecs(wordList, model):
name = []
vecs = []
for word in wordList:
word = word.replace('\n', '')
try:
if word in model: # 模型中存在该词的向量表示
name.append(word.encode('utf8').decode("utf-8"))
vecs.append(model[word])
except KeyError:
continue
a = pd.DataFrame(name, columns=['word'])
b = pd.DataFrame(np.array(vecs, dtype='float'))
return pd.concat([a, b], axis=1)
# 数据预处理操作:分词,去停用词,词性筛选
def data_prepare(text, stopkey):
l = []
# 定义选取的词性
pos = ['n', 'nz', 'v', 'vd', 'vn', 'l', 'a', 'd']
seg = jieba.posseg.cut(text) # 分词
for i in seg:
# 去重 + 去停用词 + 词性筛选
if i.word not in l and i.word\
not in stopkey and i.flag in pos:
# print i.word
l.append(i.word)
return l
# 根据数据获取候选关键词词向量
def build_words_vecs(data, stopkey, model):
idList, titleList, abstractList = data['id'], data['title'], data['abstract']
for index in range(len(idList)):
id = idList[index]
title = titleList[index]
abstract = abstractList[index]
l_ti = data_prepare(title, stopkey) # 处理标题
l_ab = data_prepare(abstract, stopkey) # 处理摘要
# 获取候选关键词的词向量
words = np.append(l_ti, l_ab) # 拼接数组元素
words = list(set(words)) # 数组元素去重,得到候选关键词列表
wordvecs = word_vecs(words, model) # 获取候选关键词的词向量表示
# 词向量写入csv文件,每个词400维
data_vecs = pd.DataFrame(wordvecs)
data_vecs.to_csv('result/vecs/wordvecs_' + str(id) + '.csv', index=False)
print ("document ", id, " well done.")
def main():
# 读取数据集
dataFile = 'data/text.csv'
data = pd.read_csv(dataFile)
# 停用词表
stopkey = [w.strip() for w in codecs.open('data/stopWord.txt', 'r', encoding='utf-8').readlines()]
# 词向量模型
inp = 'wiki.zh.text.vector'
model = gensim.models.KeyedVectors.load_word2vec_format(inp, binary=False)
build_words_vecs(data, stopkey, model)
if __name__ == '__main__':
main()
(2)编写word2vec_result.py,实现基于WordsVec词聚类的关键词提取算法
# coding=utf-8
import os
# 导入kmeans聚类算法
from sklearn.cluster import KMeans
import pandas as pd
import numpy as np
import math
# 对词向量采用K-means聚类抽取TopK关键词
def words_kmeans(data, topK):
words = data["word"] # 词汇
vecs = data.iloc[:, 1:] # 向量表示
kmeans = KMeans(n_clusters=1, random_state=10).fit(vecs)
labels = kmeans.labels_ # 类别结果标签
labels = pd.DataFrame(labels, columns=['label'])
new_df = pd.concat([labels, vecs], axis=1)
vec_center = kmeans.cluster_centers_ # 聚类中心
# 计算距离(相似性) 采用欧几里得距离(欧式距离)
distances = []
vec_words = np.array(vecs) # 候选关键词向量,dataFrame转array
vec_center = vec_center[0] # 第一个类别聚类中心,本例只有一个类别
length = len(vec_center) # 向量维度
for index in range(len(vec_words)): # 候选关键词个数
cur_wordvec = vec_words[index] # 当前词语的词向量
dis = 0 # 向量距离
for index2 in range(length):
dis += (vec_center[index2] - cur_wordvec[index2]) * \
(vec_center[index2] - cur_wordvec[index2])
dis = math.sqrt(dis)
distances.append(dis)
distances = pd.DataFrame(distances, columns=['dis'])
# 拼接词语与其对应中心点的距离
result = pd.concat([words, labels, distances], axis=1)
# 按照距离大小进行升序排序
result = result.sort_values(by="dis", ascending=True)
# 抽取排名前topK个词语作为文本关键词
wordlist = np.array(result['word'])
# 抽取前topK个词汇
word_split = [wordlist[x] for x in range(0, topK)]
word_split = " ".join(word_split)
return word_split
def main():
# 读取数据集
dataFile = 'data/text.csv'
articleData = pd.read_csv(dataFile)
ids, titles, keys = [], [], []
rootdir = "result/vecs" # 词向量文件根目录
fileList = os.listdir(rootdir) # 列出文件夹下所有的目录与文件
# 遍历文件
for i in range(len(fileList)):
filename = fileList[i]
path = os.path.join(rootdir, filename)
if os.path.isfile(path):
# 读取词向量文件数据
data = pd.read_csv(path, encoding='utf-8')
# 聚类算法得到当前文件的关键词
artile_keys = words_kmeans(data, 5)
# 根据文件名获得文章id以及标题
(shortname, extension) = os.path.splitext(filename)
t = shortname.split("_")
article_id = int(t[len(t) - 1]) # 获得文章id
# 获得文章标题
artile_tit = articleData[articleData.id ==
article_id]['title']
print(artile_tit)
print(list(artile_tit))
artile_tit = list(artile_tit)[0] # series转成字符串
ids.append(article_id)
titles.append(artile_tit)
keys.append(artile_keys.encode("utf-8").decode("utf-8"))
# 所有结果写入文件
result = pd.DataFrame({"id": ids, "title": titles, "key": keys},
columns=['id', 'title', 'key'])
result = result.sort_values(by="id", ascending=True) # 排序
result.to_csv("result/word2vec.csv", index=False,
encoding='utf_8_sig')
if __name__ == '__main__':
main()