A Comprehensive Survey of Sentiment Analysis Based on User Opinion

Abstract

In this modern era online shopping is getting a lot of attention. Thousands of reviews are available from the customers on different social media platforms which makes it difficult for the user to make a purchasing decision. For a better understanding of user opinion, sentiment analysis (also known as opinion mining) has been conducted which makes a major effect on the purchasing decision of the user. Opinion mining is defined in terms of entities, emotions, and textual relationships. User opinions on e-commerce websites or social media apps have a huge impact on product stakeholders. Over the past decades, researchers, the public sector, and the service industry are carrying out opinion mining, to eradicate and examine community sentiments and opinions. This paper presents a survey of recent studies conducted for sentiment analysis based on user opinion through machine learning techniques (focusing on supervised, semi-supervised, reinforcement, and unsupervised learning), deep learning techniques (focusing on CNN, RNN, and LSTM), and provide the background knowledge.