A Comparative Study of Different Classification Techniques for Sentiment Analysis

Sentiment analysis denotes the analysis of emotions and opinions from text. The authors also refer to sentiment analysis as opinion mining. It finds and justifies the sentiment of the person with respect to a given source of content. Social media contain vast amounts of the sentiment data in the form of product reviews, tweets, blogs, and updates on the statuses, posts, etc. Sentiment analysis of this largely generated data is very useful to express the opinion of the mass in terms of product reviews. This work is proposing a highly accurate model of sentiment analysis for reviews of products, movies, and restaurants from Amazon, IMDB, and Yelp, respectively. With the help of classifiers such as logistic regression, support vector machine, and decision tree, the authors can classify these reviews as positive or negative with higher accuracy values.