Logo Questions Linux Laravel Mysql Ubuntu Git Menu
 

Using my own corpus instead of movie_reviews corpus for Classification in NLTK

I use following code and I get it form Classification using movie review corpus in NLTK/Python

import string
from itertools import chain
from nltk.corpus import movie_reviews as mr
from nltk.corpus import stopwords
from nltk.probability import FreqDist
from nltk.classify import NaiveBayesClassifier as nbc
import nltk

stop = stopwords.words('english')
documents = [([w for w in mr.words(i) if w.lower() not in stop and w.lower() not in string.punctuation], i.split('/')[0]) for i in mr.fileids()]

word_features = FreqDist(chain(*[i for i,j in documents]))
word_features = word_features.keys()[:100]

numtrain = int(len(documents) * 90 / 100)
train_set = [({i:(i in tokens) for i in word_features}, tag) for tokens,tag in documents[:numtrain]]
test_set = [({i:(i in tokens) for i in word_features}, tag) for tokens,tag  in documents[numtrain:]]

classifier = nbc.train(train_set)
print nltk.classify.accuracy(classifier, test_set)
classifier.show_most_informative_features(5)

output:

0.655
Most Informative Features
                 bad = True              neg : pos    =      2.0 : 1.0
              script = True              neg : pos    =      1.5 : 1.0
               world = True              pos : neg    =      1.5 : 1.0
             nothing = True              neg : pos    =      1.5 : 1.0
                 bad = False             pos : neg    =      1.5 : 1.0

I want to create my own folder instead of movie_reviews in nltk, and put my own files inside it.

like image 940
ZaM Avatar asked Mar 26 '15 10:03

ZaM


1 Answers

If you have you data in exactly the same structure as the movie_review corpus in NLTK, there are two ways to "hack" your way through:

1. Put your corpus directory into where you save the nltk.data

First check where is your nltk.data saved:

>>> import nltk
>>> nltk.data.find('corpora/movie_reviews')
FileSystemPathPointer(u'/home/alvas/nltk_data/corpora/movie_reviews')

Then move your directory to where the location where nltk_data/corpora is saved:

# Let's make a test corpus like `nltk.corpus.movie_reviews`
~$ mkdir my_movie_reviews
~$ mkdir my_movie_reviews/pos
~$ mkdir my_movie_reviews/neg
~$ echo "This is a great restaurant." > my_movie_reviews/pos/1.txt
~$ echo "Had a great time at chez jerome." > my_movie_reviews/pos/2.txt
~$ echo "Food fit for the ****" > my_movie_reviews/neg/1.txt
~$ echo "Slow service." > my_movie_reviews/neg/2.txt
~$ echo "README please" > my_movie_reviews/README
# Move it to `nltk_data/corpora/`
~$ mv my_movie_reviews/ nltk_data/corpora/

In your python code:

>>> import string
>>> from nltk.corpus import LazyCorpusLoader, CategorizedPlaintextCorpusReader
>>> from nltk.corpus import stopwords
>>> my_movie_reviews = LazyCorpusLoader('my_movie_reviews', CategorizedPlaintextCorpusReader, r'(?!\.).*\.txt', cat_pattern=r'(neg|pos)/.*', encoding='ascii')
>>> mr = my_movie_reviews
>>>
>>> stop = stopwords.words('english')
>>> documents = [([w for w in mr.words(i) if w.lower() not in stop and w.lower() not in string.punctuation], i.split('/')[0]) for i in mr.fileids()]
>>> for i in documents:
...     print i
... 
([u'Food', u'fit', u'****'], u'neg')
([u'Slow', u'service'], u'neg')
([u'great', u'restaurant'], u'pos')
([u'great', u'time', u'chez', u'jerome'], u'pos')

(For more details, see https://github.com/nltk/nltk/blob/develop/nltk/corpus/util.py#L21 and https://github.com/nltk/nltk/blob/develop/nltk/corpus/init.py#L144)

2. Create your own CategorizedPlaintextCorpusReader

If you have no access to nltk.data directory and you want to use your own corpus, try this:

# Let's say that your corpus is saved on `/home/alvas/my_movie_reviews/`

>>> import string; from nltk.corpus import stopwords
>>> from nltk.corpus import CategorizedPlaintextCorpusReader
>>> mr = CategorizedPlaintextCorpusReader('/home/alvas/my_movie_reviews', r'(?!\.).*\.txt', cat_pattern=r'(neg|pos)/.*', encoding='ascii')
>>> stop = stopwords.words('english')
>>> documents = [([w for w in mr.words(i) if w.lower() not in stop and w.lower() not in string.punctuation], i.split('/')[0]) for i in mr.fileids()]
>>> 
>>> for doc in documents:
...     print doc
... 
([u'Food', u'fit', u'****'], 'neg')
([u'Slow', u'service'], 'neg')
([u'great', u'restaurant'], 'pos')
([u'great', u'time', u'chez', u'jerome'], 'pos')

Similar questions has been asked on Creating a custom categorized corpus in NLTK and Python and Using my own corpus for category classification in Python NLTK


Here's the full code that will work:

import string
from itertools import chain

from nltk.corpus import stopwords
from nltk.probability import FreqDist
from nltk.classify import NaiveBayesClassifier as nbc
from nltk.corpus import CategorizedPlaintextCorpusReader
import nltk

mydir = '/home/alvas/my_movie_reviews'

mr = CategorizedPlaintextCorpusReader(mydir, r'(?!\.).*\.txt', cat_pattern=r'(neg|pos)/.*', encoding='ascii')
stop = stopwords.words('english')
documents = [([w for w in mr.words(i) if w.lower() not in stop and w.lower() not in string.punctuation], i.split('/')[0]) for i in mr.fileids()]

word_features = FreqDist(chain(*[i for i,j in documents]))
word_features = word_features.keys()[:100]

numtrain = int(len(documents) * 90 / 100)
train_set = [({i:(i in tokens) for i in word_features}, tag) for tokens,tag in documents[:numtrain]]
test_set = [({i:(i in tokens) for i in word_features}, tag) for tokens,tag  in documents[numtrain:]]

classifier = nbc.train(train_set)
print nltk.classify.accuracy(classifier, test_set)
classifier.show_most_informative_features(5)
like image 62
alvas Avatar answered Sep 26 '22 15:09

alvas