Using data obtained from Rotten Tomatoes, we build models to classify the movie as a "fresh" or "rotten," i.e. good or bad.
We use n-gram bag of words representation and build vector models using CountVectorizer and TF-IDF Vectorizer. Naive Bayes classifier achieves a prediction accuracy of 74%, while Random Forest Classifiers achieves a prediction accuracy of 67%. Few of the top features obtained using this model include words like touching, masterpiece, superb, moving for a fresh review and words like unfortunately, lame, problem, uninspired for a rotten review.