A comprehensive survey on sentiment analysis: Challenges and future insights

Abstract

With the rise of social networks, people now express their sentiments more frequently and comfortably through their social media activities on different events, person, and every little thing surrounding them. This generates a lot of unstructured data; billions of users post tweets every day as a daily regime on Twitter itself. This has given rise to many texts classification and analysis tasks, Sentiment Analysis (SA) being one of them. Through SA, it is conferred whether the users have negative or positive orientations in their opinions; the results of this task are significantly useful for decision-makers in various fields. This paper presents various facets of SA, like the process followed in SA, levels, approaches, and sentences considered in SA. Aspects such as growth, techniques, the share of various platforms, and SA pipeline are also covered in this paper. At last, we have highlighted some major challenges in order to define future directions.