I trained a Naive Bayes bigram classifier with a Cornell University dataset of IMDB reviews, pre-classified as either positive or negative. Naives Bayes is a probablistic classifier technique that uses Bayes' Thereom where H is the class and E is the predictor.
After training the model, the program is able to analyze the text of unclassified IMDB reviews and make a binary decision, classifying a review as either positive or negative. In addition, I implemented Laplace smoothing and stop-word filtering to improve the accuracy of the classifier
Type python NaiveBayes.py data/imdb into the command line to see the 10-fold Cross Validation.