Brief analysis of the collocations of the Monty Python and the Holy Grail script.

monty-python-collocations.py

import nltk
from nltk.collocations import *
from nltk.book import *
import re

bigram_measures = nltk.collocations.BigramAssocMeasures()

# Monty Python and the Holy Grail

# Reduces tokens to words. Ignores ALL CAPS words, which are the speaker in the movie.
# Then, creates a single string of the words.
montypython_words = sum([re.findall(r'[A-Z]?[a-z]+', token) for token in text6.tokens], [])

# Finds all bigrams in the Monty Python script
# Ignore all bigrams which occur less than three times in the corpus
finder = BigramCollocationFinder.from_words(montypython_words)
finder.apply_freq_filter(3)

# Finds top 10 bigrams using the Student's t test
# Also, calculates the Student's t test for all bigrams
finder.nbest(bigram_measures.student_t, 10)
finder.score_ngrams(bigram_measures.student_t)

# Finds top 10 bigrams using Pointwise Mutual Information (PMI)
# Also, calculates the PMI for all bigrams
finder.nbest(bigram_measures.pmi, 10)
finder.score_ngrams(bigram_measures.pmi)

# Finds top 10 bigrams using the likelihood ratio
# Also, calculates the likelihood ratio for all bigrams
finder.nbest(bigram_measures.likelihood_ratio, 10)
finder.score_ngrams(bigram_measures.likelihood_ratio)

# Finds top 10 bigrams using the Pearson's Chi-squared test
# Also, calculates the Pearson's Chi-squared test for all bigrams
finder.nbest(bigram_measures.chi_sq, 10)
finder.score_ngrams(bigram_measures.chi_sq)