OPINION MINING IN TWEETS
Problem Tweets data gathered in 4 months were subjected to opinion mining and hashtag polarity examination. Solution SVG, Naïve Bayes and ID3 models were used for classification, aided by the bag-of-words feature extraction technique. Results SVM classification: 73.67% accuracy, Naïve Bayes classification: 71.01% accuracy and ID3 classification: 67.08% accuracy.