"A data mining experiment: movie reviews classification using #WEKA"
An interesting result of the experiment is that the best results have been obtained without removal of stop words and without stemming.
I have replicated the experiment with #sklearn, and confirmed it. Using count vectorizer to represent the documents and logistic regression instead of Bayesian networks the results are yet a little better. #MachineLearning #nlproc #SentimentDetection #SentimentAnalysis
http://bit.ly/2BJSn9p