The Multifaceted Concept of Context in Sentiment Analysis

Abstract

The contemporary web is about communication, collaboration, participation, and sharing. Currently, the sharing of content on the web ranges from sharing of ideas and information which includes text, photos, videos, audios, and memes to even gifs. Moreover, the language and linguistic tone of user-generated content are informal and indistinct. Analyzing explicit and clear sentiment is challenging owing to language constructs which may intensify or flip the polarity within the posts. Context-based sentiment analysis is the domain of study which deals with comprehending cues which can enhance the prediction accuracy of the generic sentiment analysis as well as facilitate fine grain analysis of varied linguistic constructs such as sarcasm, humor, or irony. This work is preliminary to understand the what, how and why of using the context in sentiment analysis. The concept of ‘context in use’ is described by exemplifying the types of context. A strength–weakness–opportunity–threat (SWOT) matrix is made to demonstrate the effectiveness of context-based sentiment analysis.