赞
踩
参考文章:自然语言处理库——NLTK_满腹的小不甘-CSDN博客
NLP 自然语言处理的开发环境搭建_村雨遥-CSDN博客_nlp开发
NLTK :: Natural Language Toolkit(官网)
手动下载并安装nltk_data_justlpf的专栏-CSDN博客
GitHub - nltk/nltk_data: NLTK Data
目录
3. 词汇规范化(Lexicon Normalization)
- Natural Language Toolkit,自然语言处理工具包,在NLP领域中,最常使用的一个Python库。
- NLTK是一个开源的项目,包含:Python模块,数据集和教程,用于NLP的研究和开发。
- NLTK由Steven Bird和Edward Loper在宾夕法尼亚大学计算机和信息科学系开发。
- NLTK包括图形演示和示例数据。其提供的教程解释了工具包支持的语言处理任务背后的基本概念。
NLTK(www.nltk.org)是在处理预料库、分类文本、分析语言结构等多项操作中最长遇到的包。其收集的大量公开数据集、模型上提供了全面、易用的接口,涵盖了分词、词性标注(Part-Of-Speech tag, POS-tag)、命名实体识别(Named Entity Recognition, NER)、句法分析(Syntactic Parse)等各项 NLP 领域的功能。
NLTK能干啥?
- 搜索文本
- 单词搜索:
- 相似词搜索;
- 相似关键词识别;
- 词汇分布图;
- 生成文本;
- 计数词汇
NLTK设计目标
- 简易性;
- 一致性;
- 可扩展性;
- 模块化;
NLTK中的语料库
- 古腾堡语料库:
gutenberg
;- 网络聊天语料库:
webtext
、nps_chat
;- 布朗语料库:
brown
;- 路透社语料库:
reuters
;- 就职演说语料库:
inaugural
;- 其他语料库;
文本语料库结构
- isolated: 独立型;
- categorized:分类型;
- overlapping:重叠型;
- temporal:暂时型;
基本语料库函数
条件频率分布
- Python安装
- NLTK系统安装
- 自动下载nltk_data一般会失败, 手动下载并配置nltk_data, 参考:手动下载并安装nltk_data_justlpf的专栏-CSDN博客
文本是由段落(Paragraph)构成的,段落是由句子(Sentence)构成的,句子是由单词构成的。切词是文本分析的第一步,它把文本段落分解为较小的实体(如单词或句子),每一个实体叫做一个Token,Token是构成句子(sentence )的单词、是段落(paragraph)的句子。NLTK能够实现句子切分和单词切分两种功能。
把段落切分成句子:
- from nltk.tokenize import sent_tokenize
-
- text="""Hello Mr. Smith, how are you doing today? The weather is great, and
- city is awesome.The sky is pinkish-blue. You shouldn't eat cardboard"""
-
- tokenized_text=sent_tokenize(text)
-
- print(tokenized_text)
- '''
- 结果:
- ['Hello Mr. Smith, how are you doing today?',
- 'The weather is great, and city is awesome.The sky is pinkish-blue.',
- "You shouldn't eat cardboard"]
- '''
句子切分成单词:
- import nltk
-
- sent = "I am almost dead this time"
- token = nltk.word_tokenize(sent)
- # 结果:token['I','am','almost','dead','this','time']
对切词的处理,需要移除标点符号和移除停用词和词汇规范化。
对每个切词调用该函数,移除字符串中的标点符号,string.punctuation包含了所有的标点符号,从切词中把这些标点符号替换为空格。
- import string
-
- """移除标点符号"""
- if __name__ == '__main__':
- # 方式一
- # s = 'abc.'
- text_list = "Hello Mr. Smith, how are you doing today? The weather is great, and city is awesome."
- text_list = text_list.translate(str.maketrans(string.punctuation, " " * len(string.punctuation))) # abc
- print("s: ", text_list)
-
-
- # 方式二
- english_punctuations = [',', '.', ':', ';', '?', '(', ')', '[', ']', '&', '!', '*', '@', '#', '$', '%']
- text_list = [word for word in text_list if word not in english_punctuations]
- print("text: ", text_list)
停用词(stopword)是文本中的噪音单词,没有任何意义,常用的英语停用词,例如:is, am, are, this, a, an, the。NLTK的语料库中有一个停用词,用户必须从切词列表中把停用词去掉。
- import nltk
- from nltk.corpus import stopwords
-
- # nltk.download('stopwords')
- # Downloading package stopwords to
- # C:\Users\Administrator\AppData\Roaming\nltk_data\corpora\stopwords.zip.
- # Unzipping the stopwords.zip
-
- """移除停用词"""
- stop_words = stopwords.words("english")
-
- if __name__ == '__main__':
- text = "Hello Mr. Smith, how are you doing today? The weather is great, and city is awesome."
-
- word_tokens = nltk.tokenize.word_tokenize(text.strip())
- filtered_word = [w for w in word_tokens if not w in stop_words]
-
- print("word_tokens: ", word_tokens)
- print("filtered_word: ", filtered_word)
- '''
- word_tokens:['Hello', 'Mr.', 'Smith', ',', 'how', 'are', 'you', 'doing', 'today', '?',
- 'The', 'weather', 'is', 'great', ',', 'and', 'city', 'is', 'awesome', '.']
- filtered_word:['Hello', 'Mr.', 'Smith', ',', 'today', '?', 'The', 'weather', 'great', ',', 'city', 'awesome', '.']
- '''
词汇规范化是指把词的各种派生形式转换为词根,在NLTK中存在两种抽取词干的方法porter和wordnet。
利用上下文语境和词性来确定相关单词的变化形式,根据词性来获取相关的词根,也叫lemma,结果是真实的单词。
基于字典的映射。nltk中要求手动注明词性,否则可能会有问题。因此一般先要分词、词性标注,再词性还原。
- from nltk.stem import WordNetLemmatizer
- lemmatizer = WordNetLemmatizer()
- lemmatizer.lemmatize('leaves')
- # 输出:'leaf'
从单词中删除词缀并返回词干,可能不是真正的单词。
- # 基于Porter词干提取算法
- from nltk.stem.porter import PorterStemmer
- porter_stemmer = PorterStemmer()
- porter_stemmer.stem(‘maximum’)
-
- # 基于Lancaster 词干提取算法
- from nltk.stem.lancaster import LancasterStemmer
- lancaster_stemmer = LancasterStemmer()
- lancaster_stemmer.stem(‘maximum’)
-
- # 基于Snowball 词干提取算法
- from nltk.stem import SnowballStemmer
- snowball_stemmer = SnowballStemmer(“english”)
- snowball_stemmer.stem(‘maximum’)
- from nltk.stem.wordnet import WordNetLemmatizer # from nltk.stem import WordNetLemmatizer
- lem = WordNetLemmatizer() # 词形还原
-
- from nltk.stem.porter import PorterStemmer # from nltk.stem import PorterStemmer
- stem = PorterStemmer() # 词干提取
-
- word = "flying"
- print("Lemmatized Word:",lem.lemmatize(word,"v"))
- print("Stemmed Word:",stem.stem(word))
- '''
- Lemmatized Word: fly
- Stemmed Word: fli
- '''
词性(POS)标记的主要目标是识别给定单词的语法组,POS标记查找句子内的关系,并为该单词分配相应的标签。
- sent = "Albert Einstein was born in Ulm, Germany in 1879."
- tokens = nltk.word_tokenize(sent)
-
- tags = nltk.pos_tag(tokens)
-
- '''
- [('Albert', 'NNP'), ('Einstein', 'NNP'), ('was', 'VBD'), ('born', 'VBN'),
- ('in', 'IN'), ('Ulm', 'NNP'), (',', ','), ('Germany', 'NNP'), ('in', 'IN'), ('1879', 'CD'), ('.', '.')]
- '''
查看一个单词的同义词集用synsets(); 它有一个参数pos,可以指定查找的词性。WordNet接口是面向语义的英语词典,类似于传统字典。它是NLTK语料库的一部分。
- import nltk
- nltk.download('wordnet') # Downloading package wordnet to C:\Users\Administrator\AppData\Roaming\nltk_data...Unzipping corpora\wordnet.zip.
-
- from nltk.corpus import wordnet
-
- word = wordnet.synsets('spectacular')
- print(word)
- # [Synset('spectacular.n.01'), Synset('dramatic.s.02'), Synset('spectacular.s.02'), Synset('outstanding.s.02')]
-
- print(word[0].definition())
- print(word[1].definition())
- print(word[2].definition())
- print(word[3].definition())
- '''
- a lavishly produced performance
- sensational in appearance or thrilling in effect
- characteristic of spectacles or drama
- having a quality that thrusts itself into attention
- '''
把切分好的词表进行词频排序(按照出现次数排序):
- all_words = nltk.FreqDist(w.lower() for w in nltk.word_tokenize( "I'm foolish foolish man" ))
- print (all_words.keys())
- all_words.plot()
dict_keys(["'m", 'man', 'i', 'foolish']):
只考虑最高频率的两个词,并且绘制累积图:
all_words.plot(
2
, cumulative
=
True
)
- #!/usr/bin/env python
- # -*- coding: utf-8 -*-
- # @Time : 2018-9-28 22:21
- # @Author : Manu
- # @Site :
- # @File : python_base.py
- # @Software: PyCharm
-
- from __future__ import division
- import nltk
- import matplotlib
- from nltk.book import *
- from nltk.util import bigrams
-
- # 单词搜索
- print('单词搜索')
- text1.concordance('boy')
- text2.concordance('friends')
-
- # 相似词搜索
- print('相似词搜索')
- text3.similar('time')
-
- #共同上下文搜索
- print('共同上下文搜索')
- text2.common_contexts(['monstrous','very'])
-
- # 词汇分布表
- print('词汇分布表')
- text4.dispersion_plot(['citizens', 'American', 'freedom', 'duties'])
-
- # 词汇计数
- print('词汇计数')
- print(len(text5))
- sorted(set(text5))
- print(len(set(text5)))
-
- # 重复词密度
- print('重复词密度')
- print(len(text8) / len(set(text8)))
-
- # 关键词密度
- print('关键词密度')
- print(text9.count('girl'))
- print(text9.count('girl') * 100 / len(text9))
-
- # 频率分布
- fdist = FreqDist(text1)
-
- vocabulary = fdist.keys()
- for i in vocabulary:
- print(i)
-
- # 高频前20
- fdist.plot(20, cumulative = True)
-
- # 低频词
- print('低频词:')
- print(fdist.hapaxes())
-
- # 词语搭配
- print('词语搭配')
- words = list(bigrams(['louder', 'words', 'speak']))
- print(words)
Copyright © 2003-2013 www.wpsshop.cn 版权所有,并保留所有权利。