Quick bigram example in Python/NLTK

example-bigrams.py

import nltk
from nltk.corpus import stopwords
from collections import Counter

word_list = []

# Set up a quick lookup table for common words like "the" and "an" so they can be excluded
stops = set(stopwords.words('english'))

# For all 18 novels in the public domain book corpus, extract all their words
[word_list.extend(nltk.corpus.gutenberg.words(f)) for f in nltk.corpus.gutenberg.fileids()]

# Filter out words that have punctuation and make everything lower-case
cleaned_words = [w.lower() for w in word_list if w.isalnum()]

# Ask NLTK to generate a list of bigrams for the word "sun", excluding 
# those words which are too common to be interesing 
sun_bigrams = [b for b in nltk.bigrams(cleaned_words) if (b[0] == 'sun' or b[1] == 'sun') \
  and b[0] not in stops and b[1] not in stops]

HI,

I am quite new to the language processing and am stuck in the bigram counting process.

I have non-financial disclosure of 110 companies for 6 years (total of 660 reports)

I have already preprocessed my files and counted Negative and Positive words based on LM dictionary (2011).

I want to calculate the frequency of bigram as well, i.e. 2 years, upcoming period etc.

Is my process right-

1. I created bigram from original files (all 660 reports)
2. I have a dictionary of around 35 bigrams
3. Check the occurrence of bigram dictionary in the files (all reports)

Are there any available codes for this kind of process?

Thank you