Given a users review, predict the stars given by the reviewer. Implemented Naive Bayes in python and used sklearn.feature_extraction.text.CountVectorizer for token counts. The models and accuracies reported are after training the models on a small subset (534k train and 133k test) of Yelp reviews. This gave 60.035% accuracy. POS+Bigrams tagging gave 63.19% accuracy. Utils.py was provided by the instructor/TAs.