With the help of Sentiment Analysis, we can easily determine the popularity of any trending topic on Social Media. The Holy Grail for many text analysis enthusiasts is Twitter data. The total number of active twitter users per month was 326 million and they create huge amount of data on a daily basis. This data can belong to any domain. At first there was only semantic approach which is based on Lexicons and used bag of words technique. Later came Machine learning which had supervised, and unsupervised approaches followed by Deep Learning with Artificial Neural Networks. This paper concentrates on implementation and comparison between Naive Bayes, Random Forest from Machine learning and Convolutional Neural Networks from Deep Learning applied on a large set of tweets each from the fields of Movies, Sports, Politics, Technology and Stock exchange respectively. The class labels considered are Positive and Negative (did not consider neutral case) and a conclusion is drawn to decide which algorithm works well for Sentiment Analysis.
Volume 11 | 04-Special Issue
Pages: 23-31