Skip to main content

Natural Language Processing with Python NLTK part 1 - Tokenizer

Natural Language Processing

Starting with the NLP articles first we will try the tokenizer in the NLTK package. Tokenizer breaks a paragraph into the relevant sub strings or sentences based on the tokenizer you used. In this I will use the Sent tokenizer, word_tokenizer and TweetTokenizer which has its specific work to do.


import nltk
from nltk.tokenize import sent_tokenize, word_tokenize, TweetTokenizer

para = "Hello there this is the blog about NLP. In this blog I have made some posts. " \
       "I can come up with new content."
tweet = "#Fun night. :) Feeling crazy #TGIF"

# tokenizing the paragraph into sentences and words
sent = sent_tokenize(para)
word = word_tokenize(para)

# printing the output
print("this paragraph has " + str(len(sent)) + " sentences and " + str(len(word)) + " words")

# print each sentence
k = 1
for i in sent:
    print("sentence " + str(k) + " = " + i)
    k += 1

# print each word
k = 1
for i in word:
    print("word " + str(k) + " = " + i)
    k += 1

# Comparing different kinds of tokenizer
print("Comparing word_tokenizer and TweetTokenizer")
print(word_tokenize(tweet))
print(TweetTokenizer().tokenize(tweet))

The Code first tokenize for sentences and words using sent_tokenizer and word_tokenizer. The output of this code will be as follows.


As you can see the sent_tokenizer has separated the para in to separate sentences and the word tokenizer into sub strings. Notice at the end I have given a comparison to Word_tokenizer and TweetTokenizer. See how smileys are taken as a one component in TweetTokenizer and two separate things in word_tokenizer.

Labels

Labels:
Location: Sri Jayawardenepura Kotte, Sri Lanka

Popular posts from this blog

Natural Language Processing with Python NLTK part 5 - Chunking and Chinking

Natural Language Processing Using regular expression modifiers we can chunk out the PoS tagged words from the earlier example. The chunking is done with regular expressions defining a chunk rule. The Chinking defines what we need to exclude from the selection. Here are list of modifiers for Python: {1,3} = for digits, u expect 1-3 counts of digits, or "places" + = match 1 or more ? = match 0 or 1 repetitions. * = match 0 or MORE repetitions $ = matches at the end of string ^ = matches start of a string | = matches either/or. Example x|y = will match either x or y [] = range, or "variance" {x} = expect to see this amount of the preceding code. {x,y} = expect to see this x-y amounts of the preceding code source: https://pythonprogramming.net/regular-expressions-regex-tutorial-python-3/ Chunking import nltk from nltk.tokenize import word_tokenize # POS tagging sent = "This will be chunked. This is for Test. World is awesome. Hello world....
Read more

Natural Language Processing with Python NLTK part 2 - Stop Words

Natural Language Processing Stop words are the words which we ignore due to the fact that they do not generate any specific meaning to the sentence. Words like the, is, at etc. can be removed to extract the meaning of the sentence more easily. So NLTK has introduced us a stop words filter we can easily use. Let's see how it works. from nltk.corpus import stopwords from nltk.tokenize import word_tokenize sent = "As you can see this is the blog of myself which is written by Anjula" w = word_tokenize(sent) # set English stop words stop_words = set(stopwords . words( 'english' )) # list of standard stop words in English print (stop_words) # making empty arrays to store stop words and others stop_words_in_sent = [] non_stop_words = [] # Loop through to get the stop words for x in w: if x not in stop_words: non_stop_words . append(x) else : stop_words_in_sent . append(x) # print result ...
Read more

Natural Language Processing with Python NLTK part 6 - Named Entity Recognition

Natural Language Processing - NER Named entities are specific reference to something. As a part of recognizing text NLTK has allowed us to used the named entity recognition and recognize certain types of entities. Those types are as follows NE Type Examples ORGANIZATION Georgia-Pacific Corp. ,  WHO PERSON Eddy Bonte ,  President Obama LOCATION Murray River ,  Mount Everest DATE June ,  2008-06-29 TIME two fifty a m ,  1:30 p.m. MONEY 175 million Canadian Dollars ,  GBP 10.40 PERCENT twenty pct ,  18.75 % FACILITY Washington Monument ,  Stonehenge GPE South East Asia ,  Midlothian Source:  http://www.nltk.org/book/ch07.html Simple example on NER: import nltk from nltk.tokenize import word_tokenize, sent_tokenize para = " America is a country. John is a name. " sent = sent_tokenize(para) for s in sent: word = word_tokenize(s) tag = nltk . pos_tag(word) n...
Read more