sentiment-anaysis-NLTK
Alogrithm:
NaiveBayesClassifier was used as the opinion classifier and was a little bit modified by using the N most frequent words.
The Data set has been split into two:
1: Training Set - Containing 90% of the data, which was used to train the data.
2: Testing Set - Containing 10% of the data, which was used to test the data.
Analysis of movie_reviews corpora :
Accuracy on modified NaiveBayesClassifier algorithm : 86.5
Accuracy with algorithm including bigrams : 85.5