Approaches of Sentiment Analysis: A Review

Abstract

The social media websites are full of information related to the public’s opinions, and emotions in different aspects such as elections, products, movies, disaster, and company brand, etc. The opinion of people is very useful in commercializing a business. Sentiment analysis is used to determine the attitude of the writer on a topic as positive or negative. Twitter is a widely used micro-blogging site for expressing opinions. In this paper, the approaches used by different researchers for sentiment analysis is discussed and based on the accuracy, precision, recall, and f1 score, it is discovered that the lexicon-based approach is widely used in NLP and Naïve Bayes and SVM are highly effective and accurate classifiers used in Machine Learning. This paper also presents the deep learning approach and the major challenges faced by sentiment analysis.