A systematic literature review on sentiment analysis techniques, challenges, and future trends

In the last ten years, social media site mining, including Twitter, Facebook, Instagram, and all other websites, has become a popular area of study. In the world of digital media today, people are getting more and more vocal because they love to share their opinions. User-generated material abounds on social media platforms and apps such as Facebook, WhatsApp, and Twitter, providing wealthy content to gather sentiments. Comments are another way that the most active and social voices can make themselves heard, which in turn gives us a window into their feelings. Sentiment analysis is the task of comprehending the emotions and opinions conveyed in text and other media. It has various applications in domains along with social media such as e-commerce, health, politics, and marketing. This paper delineates the generic process of sentiment analysis and reviews the main methods, challenges, and trends in this field. The main goal of this survey paper is to survey the current state-of-the-art research works on sentiment analysis (SA) techniques and related fields and to compare the performance of different deep learning models for sentiment polarity. The paper also discusses the recent studies that have employed machine learning, deep learning, and hybrid models to address sentiment polarity problems, which is the categorization of text into positive, negative, or neutral sentiments. The paper evaluates the results of different models on a series of datasets. The paper's main contributions are the elaborate classifications of numerous recent articles and the demonstration of the recent research directions in sentiment analysis and its related fields. The paper aims to provide a comprehensive overview of SA techniques with succinct details.