How to get n-gram collocations and association in

问题:

In this documentation, there is example using nltk.collocations.BigramAssocMeasures(), BigramCollocationFinder,nltk.collocations.TrigramAssocMeasures(), and TrigramCollocationFinder.

There is example method find nbest based on pmi for bigram and trigram. example:

finder = BigramCollocationFinder.from_words(
...     nltk.corpus.genesis.words('english-web.txt'))
>>> finder.nbest(bigram_measures.pmi, 10)

I know that BigramCollocationFinder and TrigramCollocationFinder inherit from AbstractCollocationFinder. While BigramAssocMeasures() and TrigramAssocMeasures() inherit from NgramAssocMeasures.

How can I use the methods(e.g. nbest()) in AbstractCollocationFinder and NgramAssocMeasures for 4-gram, 5-gram, 6-gram, ...., n-gram (like using bigram and trigram easily)?

Should I create class which inherit AbstractCollocationFinder?

Thanks.

回答1:

Edited

The current NLTK has a hardcoder function for up to QuadCollocationFinder but the reasoning for why you cannot simply create an NgramCollocationFinder still stands, you would have to radically change the formulas in the from_words() function for different order of ngram.

---

Short answer, no you cannot simply create an AbstractCollocationFinder (ACF) to call the nbest() function if you want to find collocations beyond 2- and 3-grams.

It's because of the difference in the from_words() for different ngrams. You see that only the subclass of ACF (i.e. BigramCF and TrigramCF) have the from_words() function.

>>> finder = BigramCollocationFinder.from_words(nltk.corpus.genesis.words('english-web.txt'))
>>> finder = AbstractCollocationFinder.from_words(nltk.corpus.genesis.words('english-web.txt',5))
Traceback (most recent call last):
  File "<stdin>", line 1, in <module>
AttributeError: type object 'AbstractCollocationFinder' has no attribute 'from_words'

So given this from_words() in TrigramCF:

from nltk.probability import FreqDist
@classmethod
def from_words(cls, words):
    wfd, wildfd, bfd, tfd = (FreqDist(),)*4

    for w1,w2,w3 in ingrams(words,3,pad_right=True):
      wfd.inc(w1)

      if w2 is None:
        continue
      bfd.inc((w1,w2))

      if w3 is None:
        continue
      wildfd.inc((w1,w3))
      tfd.inc((w1,w2,w3))

    return cls(wfd, bfd, wildfd, tfd)

You could somehow hack it and try to hardcode for a 4-gram association finder as such:

@classmethod
def from_words(cls, words):
    wfd, wildfd = (FreqDist(),)*2
    bfd, tfd ,fofd = (FreqDist(),)*3

    for w1,w2,w3,w4,w5 in ingrams(words,5,pad_right=True):
      wfd.inc(w1)

      if w2 is None:
        continue
      bfd.inc((w1,w2))

      if w3 is None:
        continue
      wildfd.inc((w1,w3))
      tfd.inc((w1,w2,w3))

      if w4 is None:
        continue
      wildfd.inc((w1,w4))
      wildfd.inc((w2,w4))
      wildfd.inc((w3,w4))
      wildfd.inc((w1,w3))
      wildfd.inc((w2,w3))
      wildfd.inc((w1,w2))
      ffd.inc((w1,w2,w3,w4))

    return cls(wfd, bfd, wildfd, tfd, ffd)

Then you would also have to change whichever part of the code that uses cls returned from the from_words respectively.

So you have to ask what is the ultimate purpose of finding the collocations?

• If you're looking at retreiving words within collocations of larger than 2 or 3grams windows then you pretty much end up with a lot of noise in your word retrieval.

• If you're going to build a model base on a collocation mode using 2 or 3grams windows then you will also face sparsity problems.

回答2:

If you want to find the grams beyond 2 or 3 grams you can use scikit package and Freqdist function to get the count for these grams. I tried doing this with nltk.collocations, but I dont think we can find out more than 3-grams score into it. So I rather decided to go with count of grams. I hope this can help u a little bit. Thankz

here is the code

from sklearn.metrics.pairwise import cosine_similarity
from sklearn.feature_extraction.text import CountVectorizer
from nltk.collocations import *
from nltk.probability import FreqDist
import nltk

query = "This document gives a very short introduction to machine learning problems"
vect = CountVectorizer(ngram_range=(1,4))
analyzer = vect.build_analyzer()
listNgramQuery = analyzer(query)
listNgramQuery.reverse()
print "listNgramQuery=", listNgramQuery
NgramQueryWeights = nltk.FreqDist(listNgramQuery)
print "\nNgramQueryWeights=", NgramQueryWeights

This will give output as

listNgramQuery= [u'to machine learning problems', u'introduction to machine learning', u'short introduction to machine', u'very short introduction to', u'gives very short introduction', u'document gives very short', u'this document gives very', u'machine learning problems', u'to machine learning', u'introduction to machine', u'short introduction to', u'very short introduction', u'gives very short', u'document gives very', u'this document gives', u'learning problems', u'machine learning', u'to machine', u'introduction to', u'short introduction', u'very short', u'gives very', u'document gives', u'this document', u'problems', u'learning', u'machine', u'to', u'introduction', u'short', u'very', u'gives', u'document', u'this']

NgramQueryWeights= <FreqDist: u'document': 1, u'document gives': 1, u'document gives very': 1, u'document gives very short': 1, u'gives': 1, u'gives very': 1, u'gives very short': 1, u'gives very short introduction': 1, u'introduction': 1, u'introduction to': 1, ...>