Prerequisite: Introduction to NLP, Text Preprocessing in Python | Set 1
In the previous post, we saw the basic preprocessing steps when working with textual data. In this article, we will look at some more advanced text preprocessing techniques. We can use these techniques to gain more insights into the data that we have.
Let’s import the necessary libraries.
# import the necessary libraries import nltk import string import re |
Part of Speech Tagging:
The part of speech explains how a word is used in a sentence. In a sentence, a word can have different contexts and semantic meanings. The basic natural language processing models like bag-of-words fail to identify these relations between words. Hence, we use part of speech tagging to mark a word to its part of speech tag based on its context in the data. It is also used to extract relationships between words.
from nltk.tokenize import word_tokenize from nltk import pos_tag # convert text into word_tokens with their tags def pos_tagging(text): word_tokens = word_tokenize(text) return pos_tag(word_tokens) pos_tagging( 'You just gave me a scare' ) |
Example:
Input: ‘You just gave me a scare’
Output: [(‘You’, ‘PRP’), (‘just’, ‘RB’), (‘gave’, ‘VBD’), (‘me’, ‘PRP’),
(‘a’, ‘DT’), (‘scare’, ‘NN’)]
In the given example, PRP stands for personal pronoun, RB for adverb, VBD for verb past tense, DT for determiner and NN for noun. We can get the details of all the part of speech tags using the Penn Treebank tagset.
# download the tagset nltk.download( 'tagsets' ) # extract information about the tag nltk. help .upenn_tagset( 'NN' ) |
Example:
Input: ‘NN’
Output: NN: noun, common, singular or mass
common-carrier cabbage knuckle-duster Casino afghan shed thermostat
investment slide humour falloff slick wind hyena override subhumanity
machinist …
Chunking:
Chunking is the process of extracting phrases from unstructured text and more structure to it. It is also known as shallow parsing. It is done on top of Part of Speech tagging. It groups word into “chunks”, mainly of noun phrases. Chunking is done using regular expressions.
from nltk.tokenize import word_tokenize from nltk import pos_tag # define chunking function with text and regular # expression representing grammar as parameter def chunking(text, grammar): word_tokens = word_tokenize(text) # label words with part of speech word_pos = pos_tag(word_tokens) # create a chunk parser using grammar chunkParser = nltk.RegexpParser(grammar) # test it on the list of word tokens with tagged pos tree = chunkParser.parse(word_pos) for subtree in tree.subtrees(): print (subtree) tree.draw() sentence = 'the little yellow bird is flying in the sky' grammar = "NP: {<DT>?<JJ>*<NN>}" chunking(sentence, grammar) |
In the given example, grammar, which is defined using a simple regular expression rule. This rule says that an NP (Noun Phrase) chunk should be formed whenever the chunker finds an optional determiner (DT) followed by any number of adjectives (JJ) and then a noun (NN).
Libraries like spaCy and Textblob are more suited for chunking.
Example:
Input: ‘the little yellow bird is flying in the sky’
Output:
(S
(NP the/DT little/JJ yellow/JJ bird/NN)
is/VBZ
flying/VBG
in/IN
(NP the/DT sky/NN))
(NP the/DT little/JJ yellow/JJ bird/NN)
(NP the/DT sky/NN)
Named Entity Recognition:
Named Entity Recognition is used to extract information from unstructured text. It is used to classify entities present in a text into categories like a person, organization, event, places, etc. It gives us detailed knowledge about the text and the relationships between the different entities.
from nltk.tokenize import word_tokenize from nltk import pos_tag, ne_chunk def named_entity_recognition(text): # tokenize the text word_tokens = word_tokenize(text) # part of speech tagging of words word_pos = pos_tag(word_tokens) # tree of word entities print (ne_chunk(word_pos)) text = 'Bill works for Lazyroar so he went to Delhi for a meetup.' named_entity_recognition(text) |
Example:
Input: ‘Bill works for Lazyroar so he went to Delhi for a meetup.’
Output:
(S
(PERSON Bill/NNP)
works/VBZ
for/IN
(ORGANIZATION Lazyroar/NNP)
so/RB
he/PRP
went/VBD
to/TO
(GPE Delhi/NNP)
for/IN
a/DT
meetup/NN
./.)
<!–
–>
Please Login to comment…