Naive Bayes and logistic regression algorithms were used in order to classify reviews polarity (positive/negative). Also, word2vec analysis was performed.
Cornell Movie Review dataset was used. Based on Python 2.7 (3.5 also works with minimum changes), Tensor Flow 1.0