Sentiment analysis researches story narrated by topic modeling approach

Abstract

The wild growth of user-generated content like websites, social media, and mobile apps, conducts individuals to create enormous masses of opinions and reviews about products, services, and every day events. Sentiment analysis (SA) embraces a powerful tool for businesses and researchers to explore and study community attitudes, interpretations and, insightful consequences for decision support. This paper brings forward a comprehensive study about main research topics, research trends, and comparisons of research topics” in the field of “sentiment analysis” through “social media” using topic modeling, in specific LDA. The findings of this paper prove that “machine learning” methods are among the most important topics the studies worked on in recent years. Also, various social media platforms such as “Twitter, Facebook, YouTube, and blog” are the SA infrastructures. Among the applications, transportation, spam detection, and decision making are important in terms of the normalized frequency. Finally, findings verify the concept “service improvement by sentiment analysis” indicates the important topic which concentrates quality improvement of firm's service through analysis of customer reviews and it permits researchers and practitioners and also managers have better visions about the hot era of “sentiment analysis”. • A comprehensive study about main research topics, research trends, and comparisons of research topics” in the field of “sentiment analysis”. • Latent Dirichlet Allocation (LDA), a probabilistic topic modeling is used to discover latent topics from a large volume of data. • This research proves that “machine learning” methods are the most important topic. • Among the applications, transportation, spam detection, and decision making are important in terms of the normalized frequency.