This is a course final project of processing restaurant review data from Yelp.
There are mainly two function of this project:
1. Find the food-related keyword from Yelp review or from a specific restaurant's review
2. Generate a review based rating to each restaurant using Sentimental analysis and compare with the true rating.
Part 0: Tokenize the reviews
Part 1: Use nltk to find the unigram, bigram and trigram collocations
Part 2: Use tf-idf to find the collocations that are specific to each document (business)
Part 3: Find food-related collocations
Method 1: wordnet
Method 2: word2vec
Using textblob sentiment function and generate the sentimental score of reviews based on polarity and subjectivity.