A Brief Survey: Features and Techniques Used for Sentiment Analysis

Today, people are more passionate in using websites, social media, and e-shopping. They are also more eager to express and share their opinions or feedbacks on web regarding day-to-day activities and global issues. But most of the reviewers are not genuine in giving reviews and opinions. Some people may create false reviews to promote or disparage products and services. This practice of making fake, untruthful, or deceptive reviews is known as opinion spam. Over the past few decades, research communities, academia, public, and industries are meticulously working on sentiment analysis (opinion mining) to extract and predict the authentic interests and usage requirements of customers. Using natural language processing, meaningful features can be extracted from the text and are possible to conduct review spam detection using various machine learning techniques. In this paper, we emphasized and analyzed the various feature selection methods in sentiment analysis by studying the various papers of different authors.