Sentiment Analysis using Python's NLTK
The purpose of this implementation is to classify a piece of information (product reviews in this case) as either positive or negative. Data sets can be found here
About 700 positive and 700 negative reviews are used to train the classifier. Naive Bayes is the classification algorithm used in this implementation. The Naive Bayes algorithm works on the principle that every feature contributes independently to the classification.
Given a set of hypotheses H, and a classifier X,
P(X|H) = P(X)P(H|X) / P(H)
MongoDB is used to store the reviews